CN114343555A - A colposcopy artificial intelligence-assisted quality control management method and device - Google Patents

Abstract

Translated fromChinese

本申请提供了一种阴道镜人工智能辅助质控管理方法和装置，获取按照第一时序采集的阴道镜检查的过程中的多模态阴道镜图像；根据所有的多模态阴道镜图像的变化情况，确定拟诊结果，按照预先设定的评估信息，对拟诊结果进行评分，得到评估总分；获取转诊信息；根据转诊信息和评估总分，确定预测结果。按照第一时序采集阴道镜检查过程中的多模态阴道镜图像，根据时序变化的多模态阴道镜图像的关联特征共同确定拟诊结果，根据该拟诊结果能够找出重点关注的信息；本方案采用评分式评估，利用预先设定的评估信息对拟诊结果进行评分，得到评估总分；将该评估总分关联转诊信息共同对当前的阴道镜诊断进行预测，让评估更加客观，提高了预测结果的可靠性。

The present application provides a colposcopy artificial intelligence-assisted quality control management method and device, which acquires multimodal colposcopy images in the process of colposcopy collected according to a first sequence; According to the pre-set evaluation information, the proposed diagnosis result is scored to obtain the total evaluation score; the referral information is obtained; the prediction result is determined according to the referral information and the total evaluation score. Collect the multimodal colposcopy images in the colposcopy process according to the first time sequence, jointly determine the proposed diagnosis result according to the correlation characteristics of the multimodal colposcopy images that change in the time sequence, and find out the information of key concern according to the proposed diagnosis result; This program adopts scoring evaluation, using pre-set evaluation information to score the results of the proposed diagnosis to obtain the total evaluation score; the total evaluation score is associated with the referral information to predict the current colposcopy diagnosis, making the evaluation more objective, Improved reliability of forecast results.

Description

Translated fromChinese

一种阴道镜人工智能辅助质控管理方法和装置A colposcopy artificial intelligence-assisted quality control management method and device

技术领域technical field

本发明涉及阴道镜质控管理的技术领域，尤其是涉及一种阴道镜人工智能辅助质控管理方法和装置。The invention relates to the technical field of colposcopy quality control management, in particular to a colposcopy artificial intelligence-assisted quality control management method and device.

背景技术Background technique

阴道镜是一种妇科临床诊断仪器，是妇科内窥镜之一。适用于各种宫颈疾病及生殖器病变的诊断，也是男女性疾病早期诊断的重要方式。阴道镜能将观测到的图像放大10-60倍，发现肉眼不能发现的微小病变。借着这种放大效果，医生可以清楚地看到子宫颈表皮和生殖器表皮上极其微小的病灶细节，有助于提高判断宫颈、生殖器等病变的准确率，为疾病的早期诊断提供依据，使患者提前得到有效的治疗，使疾病的治愈率极大提高。Colposcopy is a gynecological clinical diagnostic instrument and one of gynecological endoscopes. It is suitable for the diagnosis of various cervical diseases and genital lesions, and is also an important method for the early diagnosis of male and female diseases. Colposcopy can magnify the observed image by 10-60 times and find tiny lesions that cannot be seen by the naked eye. With this magnifying effect, doctors can clearly see the details of extremely small lesions on the epidermis of the cervix and genitals, which helps to improve the accuracy of judging cervical and genital lesions, and provides a basis for early diagnosis of the disease. Early effective treatment can greatly improve the cure rate of the disease.

阴道镜检查为非接触检查，借助专用窥器、子宫颈管扩张器来暴露检查部位，阴道镜与宫颈口的工作距离为200-300mm，能够避免交叉感染，适合快速的体检普查，阴道镜检查包括醋白检查，或者，包括醋白检查、补碘检查。Colposcopy is a non-contact examination. The examination site is exposed with the help of a special speculum and cervical dilator. The working distance between the colposcope and the cervix is 200-300mm, which can avoid cross-infection and is suitable for rapid physical examination and colposcopy. Including vinegar white inspection, or, including vinegar white inspection, iodine supplementation inspection.

相关技术的阴道镜检查方式是追踪阴道镜检查过程中阴道镜图像的特征间的关联变化直接进行诊断，仅针对特征的关联变化进行诊断容易造成诊断结果的不客观性，造成诊断的结果准确性低的问题。The colposcopy method of the related art is to track the correlation changes between the features of the colposcopy image during the process of colposcopy to make a diagnosis directly. Diagnosis only based on the correlation changes of the features will easily lead to the non-objectivity of the diagnosis results and the accuracy of the diagnosis results. low problem.

发明内容SUMMARY OF THE INVENTION

本发明的发明目的一是提供一种阴道镜人工智能辅助质控管理方法，通过以下技术方案得以实现的：One object of the invention of the present invention is to provide a colposcopy artificial intelligence-assisted quality control management method, which is achieved through the following technical solutions:

一种阴道镜人工智能辅助质控管理方法，包括：A colposcopy artificial intelligence-assisted quality control management method, comprising:

获取按照第一时序采集的阴道镜检查的过程中的多模态阴道镜图像，其中，所述阴道镜检查包括醋白检查，或，醋白检查和补碘检查；acquiring a multimodal colposcopy image during a colposcopy examination collected according to a first time sequence, wherein the colposcopy examination includes acetowhite examination, or acetowhite examination and iodine supplementation examination;

根据所有的所述多模态阴道镜图像的变化情况，确定拟诊结果，其中，所述拟诊结果包括：转化区病变大小、出现白上皮、醋白上皮表面和边界、血管、转化区类型、其它信息的性状，所述其他信息的性状包括可见持续性出血，可见易脆组织，可见白斑涂醋酸前存在是否可见；According to the changes of all the multimodal colposcopy images, the proposed diagnosis results are determined, wherein the proposed diagnosis results include: the size of the lesion in the transformation zone, the appearance of white epithelium, the surface and border of the acetowhite epithelium, blood vessels, and the type of the transformation zone , The traits of other information, the traits of the other information include visible persistent bleeding, visible brittle tissue, visible vitiligo before application of acetic acid or not;

当所述阴道镜检查包括醋白检查和补碘检查时，所述拟诊结果还包括碘染色的性状；When the colposcopy examination includes acetowhite examination and iodine supplementation examination, the proposed diagnosis result also includes the characteristics of iodine staining;

按照预先设定的评估信息，对所述拟诊结果进行评分，得到评估总分；According to the pre-set evaluation information, the proposed diagnosis result is scored to obtain the total evaluation score;

获取转诊信息；其中，所述转诊信息包括：HPV检测信息、细胞学信息；Obtain referral information; wherein, the referral information includes: HPV detection information and cytology information;

根据所述转诊信息和所述评估总分，确定预测结果，其中所述预测结果包括病变类型和病变部位活检提示。According to the referral information and the total evaluation score, a prediction result is determined, wherein the prediction result includes the lesion type and the biopsy prompt of the lesion site.

可见，本方案按照第一时序采集阴道镜检查过程中的多模态阴道镜图像，保证采集的质控管理，即使是经验不足的医疗工作者也可以执行，然后，根据获取的时序变化的多模态阴道镜图像的关联特征共同确定拟诊结果，根据该拟诊结果能够找出重点关注的信息；基于人工智能，利用预先设定的评估信息对拟诊结果进行评分，得到评估总分；将该评估总分关联转诊信息共同对当前的阴道镜诊断进行预测，让评估更加客观，提高了预测结果的可靠性。It can be seen that this scheme collects the multimodal colposcopy images during the colposcopy process according to the first time sequence to ensure the quality control management of the collection, and even inexperienced medical workers can perform it. The correlation characteristics of the modal colposcopy images jointly determine the suspected diagnosis results, and the key information can be found according to the suspected diagnosis results; based on artificial intelligence, the pre-set evaluation information is used to score the suspected diagnosis results, and the total evaluation score is obtained; The total evaluation score is associated with referral information to predict the current colposcopy diagnosis, which makes the evaluation more objective and improves the reliability of the predicted results.

优选的，阴道镜人工智能辅助质控管理方法，还包括：Preferably, the colposcopy artificial intelligence-assisted quality control management method further includes:

当患者对所述病变部位进行活检后，获取所述病变部位对应的活检结果；After the patient performs a biopsy on the lesion site, obtain the biopsy result corresponding to the lesion site;

利用所述活检结果对所述预测结果进行复检，得到复检结果。The predicted result is re-examined using the biopsy result to obtain a re-examination result.

可见，本方案患者会对病变部位进行活检，并将活检结果上传至阴道镜人工智能辅助质控管理系统，在阴道镜人工智能辅助质控管理系统接收到该活检结果后，利用活检结果对预测结果进行自动复检，确保复检结果真实有效。It can be seen that the patients in this scheme will biopsy the lesion and upload the biopsy results to the colposcopy AI-assisted quality control management system. After the colposcopy AI-assisted quality control management system receives the biopsy results, the biopsy results are used to predict the results. The results are automatically re-examined to ensure that the re-examination results are true and valid.

优选的，还包括：Preferably, it also includes:

将所述预测结果、所有多模态阴道镜图像以及所述转诊信息上传至云端服务器，以使所述云端服务器对应的医疗工作者进行人工分析，得到分析结果；uploading the prediction result, all the multimodal colposcopy images and the referral information to the cloud server, so that the medical worker corresponding to the cloud server can perform manual analysis to obtain the analysis result;

从所述云端服务器获取所述分析结果；Obtain the analysis result from the cloud server;

根据所述分析结果对所述预测结果进行修正。The prediction result is revised according to the analysis result.

可见，本方案将相关数据传输至云端服务器，以便至少一位有经验的医疗工作者再次分析，得到阴道镜的分析结果，再反馈给云端服务器，实现多个医疗工作者同时会诊，提高诊断的准确性。It can be seen that this scheme transmits the relevant data to the cloud server, so that at least one experienced medical worker can analyze it again, get the analysis result of the colposcopy, and then feed it back to the cloud server, so as to realize the simultaneous consultation of multiple medical workers, and improve the diagnostic accuracy. accuracy.

优选的，所述根据所有的所述多模态阴道镜图像的变化情况，确定拟诊结果，包括：在显示界面按照第一时序显示所有的所述多模态阴道镜图像，以便当前医疗工作者根据所有的所述多模态阴道镜图像的变化情况输入拟诊结果；Preferably, determining the results of the suspected diagnosis according to the changes of all the multimodal colposcopy images includes: displaying all the multimodal colposcopy images on the display interface according to a first sequence, so as to facilitate the current medical work According to the changes of all the multimodal colposcopy images, the person inputs the simulated diagnosis results;

获取所述用户确定的所述拟诊结果；obtaining the proposed diagnosis result determined by the user;

或，or,

按照第一时序发送所有的所述多模态阴道镜图像至云端服务器，以便所述云端服务器对所述多模态阴道镜图像进行自动特征识别，并根据特征的变化情况，生成所述拟诊结果；Send all the multimodal colposcopy images to the cloud server according to the first sequence, so that the cloud server can automatically identify the features of the multimodal colposcopy images, and generate the simulated diagnosis according to the changes of the features. result;

获取所述云端服务器发送的所述拟诊结果。Obtain the proposed diagnosis result sent by the cloud server.

可见，本方案提供两种确定拟诊结果的手段，医疗工作者确定能够提高准确性，而云端服务器智能识别能够提高拟诊结果的获取效率。It can be seen that this solution provides two methods for determining the results of the proposed diagnosis. The medical workers can determine that the accuracy can be improved, and the intelligent identification of the cloud server can improve the efficiency of obtaining the results of the proposed diagnosis.

优选的，所述获取按照第一时序采集的阴道镜检查的过程中的多模态阴道镜图像之前，还包括：Preferably, before the acquiring the multimodal colposcopy image during the colposcopy process collected according to the first time sequence, the method further includes:

获取镜头参数下的阴道镜图像的清晰度，所述镜头参数包括镜头位置，镜头的各光源的照度、均匀值以及色温；Obtaining the clarity of the colposcopy image under lens parameters, the lens parameters including the lens position, the illuminance, uniform value and color temperature of each light source of the lens;

发送镜头参数调整指令至阴道镜镜头，以使所述阴道镜镜头调整镜头参数，使各光源下的所述阴道镜图像的清晰度至设定清晰度；Sending a lens parameter adjustment instruction to the colposcope lens, so that the colposcope lens adjusts the lens parameters, so that the clarity of the colposcope image under each light source reaches the set clarity;

确定设定清晰度的阴道镜图像是否存在异物；Determine the presence or absence of foreign bodies in colposcopy images of set clarity;

若是，发送提示信息至显示设备，提示信息用于提示医护人员去除异物；If so, send a prompt message to the display device, and the prompt message is used to prompt the medical staff to remove the foreign body;

若否，则按照校准后的镜头参数控制阴道镜镜头进行多模态阴道图像采集。If not, the colposcope lens is controlled according to the calibrated lens parameters for multimodal vaginal image acquisition.

可见，本方案通过对镜头参数进行质控管理，尽量减少了图像采集不清楚的问题，同时如果当前宫颈对应的阴道镜图像存在异物时，也能够及时提示医护人员去除，使宫颈组织暴露出来，最终保证了后续采集的多模态阴道镜图像清楚且无异物。It can be seen that this scheme minimizes the problem of unclear image acquisition through quality control management of lens parameters. At the same time, if there is a foreign body in the colposcopy image corresponding to the current cervix, it can also prompt the medical staff to remove it in time to expose the cervical tissue. Ultimately, the multimodal colposcopy images collected subsequently are guaranteed to be clear and free of foreign objects.

优选的，还包括：Preferably, it also includes:

当宫颈部位涂覆生理盐水后，获取按照第二时序采集的多张生理盐水涂覆后的生理盐水阴道镜图像；After the cervix is coated with the physiological saline, acquiring a plurality of saline colposcopy images after the physiological saline coating collected according to the second time sequence;

对所有所述生理盐水阴道镜图像进行特征识别，确定是否具有可疑癌特征；所述可疑癌特征包括：肿瘤或新生肿物、外生型病变、脆性血管、表面不规则、坏死、溃疡；Perform feature recognition on all the saline colposcopy images to determine whether there are suspicious features of cancer; the suspicious features of cancer include: tumors or new tumors, exophytic lesions, fragile blood vessels, surface irregularities, necrosis, and ulcers;

若不具有，则在显示界面显示待评估信息，待评估信息包括：宫颈上皮雌激素作用影响、绝经后上皮萎缩性改变，检查前是否做过雌激素治疗、妊娠新改变存在、阴道镜下确定急性感染存在、镜下能充分观察到整个宫颈和阴道上段、确认是否接受过治疗或镜下可见治疗后改变、上皮纤维化或镜下不同程度可见；If not, the information to be evaluated will be displayed on the display interface. The information to be evaluated includes: the effect of estrogen on the cervical epithelium, postmenopausal epithelial atrophic changes, whether estrogen therapy has been done before the examination, the existence of new changes in pregnancy, and colposcopy confirmation. Acute infection exists, the entire cervix and upper vaginal segment can be fully observed under the microscope, whether it has been treated or the changes after treatment can be seen under the microscope, epithelial fibrosis or different degrees of visible under the microscope;

接收到当前医疗工作者从所述待评估信息中选择的已评估信息；receiving assessed information selected by the current healthcare worker from the information to be assessed;

根据已评估信息确定阴道镜检查条件是否充分，若是，则执行所述获取按照第一时序采集的阴道镜检查的过程中的多模态阴道镜图像的步骤。It is determined whether the colposcopy conditions are sufficient according to the evaluated information, and if so, the step of acquiring the multimodal colposcopy images in the process of colposcopy acquired according to the first time sequence is performed.

可见，本方案在进行醋酸检查时需要进行涂醋酸前的评估，主要是根据按照第二时序采集到的生理盐水阴道镜图像的特征确定是否具有可疑癌特征，如果不具有则需要医疗工作者填写用于确认阴道镜检查条件是否充分的评估信息，只有充分后，才能进行阴道镜检查，避免了无效检查的发生。It can be seen that the evaluation before acetic acid is required for the acetic acid examination in this scheme, mainly based on the characteristics of the saline colposcopy images collected according to the second time series to determine whether there are suspicious cancer characteristics, if not, it needs to be filled in by medical workers. The evaluation information used to confirm whether the conditions of colposcopy are sufficient, and only after sufficient information can the colposcopy be performed, avoiding the occurrence of invalid examinations.

优选的，还包括：Preferably, it also includes:

实时检测是否具有与所述阴道镜人工智能辅助质控管理的当前阶段相匹配的提示信息，若具有，则在显示界面上显示所述提示信息。Whether there is prompt information matching the current stage of the colposcopy artificial intelligence-assisted quality control management is detected in real time, and if so, the prompt information is displayed on the display interface.

可见，本方案通过对每阶段设置提示信息，以便人员能够根据提示信息执行对应的操作，或者了解当前所处的阶段，保证能够针对影响阴道镜产品质控管理的全环节因素和检查流程，更全面地规范操作流程要求，进行引导提示，保障采图质量，降低人为失误风险，提升基层诊疗水平。It can be seen that this scheme sets up prompt information for each stage, so that personnel can perform corresponding operations according to the prompt information, or understand the current stage, so as to ensure that the whole-link factors and inspection process that affect the quality control management of colposcopy products can be improved. Comprehensively standardize the operation process requirements, provide guidance and prompts, ensure the quality of maps, reduce the risk of human error, and improve the level of primary care.

优选的，所述根据所述转诊信息和所述评估总分，确定预测结果，其中所述预测结果包括病变类型和病变部位活检提示，包括：Preferably, the prediction result is determined according to the referral information and the evaluation total score, wherein the prediction result includes the lesion type and the biopsy prompt of the lesion site, including:

确定所述评估总分所处的评估范围；determining the assessment range within which the overall assessment score falls;

根据所述转诊信息和所述评估范围，确定预测结果。According to the referral information and the evaluation range, a prediction result is determined.

可见，本方案根据评估总分以及相应的转诊信息确定预测结果，该对应关系是根据大量的样本数据得到的，得到预测结果的方法简单，准确性高。It can be seen that this scheme determines the prediction result according to the total evaluation score and the corresponding referral information. The corresponding relationship is obtained from a large number of sample data. The method for obtaining the prediction result is simple and has high accuracy.

优选的，所述获取按照第一时序采集的阴道镜检查的过程中的多模态阴道镜图像，包括：Preferably, the obtaining of the multimodal colposcopy images during the colposcopy process collected according to the first time sequence includes:

获取按照第一时序采集的阴道镜检查的过程中的多模态阴道镜图像；acquiring a multimodal colposcopy image in the process of colposcopy that is collected according to the first time sequence;

判断当前时刻是否存在多模态阴道镜图像的漏检情况；Determine whether there is a missed detection of multimodal colposcopy images at the current moment;

若存在当确定当前所述多模态阴道镜图像为模糊图像，则进行漏检提示，并控制图像采集设备进行再次采集，得到补充多模态阴道镜图像；If there is, when it is determined that the current multimodal colposcopy image is a blurred image, a missed detection prompt is performed, and the image acquisition device is controlled to collect again to obtain a supplementary multimodal colposcopy image;

根据所述标记将所述模糊图像替换成所述补充多模态阴道镜图像。The blurred image is replaced with the supplemental multimodal colposcopy image according to the indicia.

可见，本方案能够及时确定漏检情况，并进行漏检提示和相应的图像替换，保证了多模态阴道镜图像的准确，进而提高了拟诊结果确定的准确性。It can be seen that this scheme can timely determine the missed detection situation, and perform missed detection prompts and corresponding image replacement, which ensures the accuracy of multimodal colposcopy images, thereby improving the accuracy of the determination of the proposed diagnosis results.

本发明的发明目的二是提供一种阴道镜人工智能辅助质控管理装置，通过以下技术方案得以实现的：The second purpose of the invention of the present invention is to provide a colposcopy artificial intelligence-assisted quality control management device, which is achieved through the following technical solutions:

一种阴道镜人工智能辅助质控管理装置，包括：A colposcopy artificial intelligence-assisted quality control management device, comprising:

多模态阴道镜图像获取模块，用于获取按照第一时序采集的阴道镜检查的过程中的多模态阴道镜图像，其中，所述阴道镜检查包括醋白检查，或，醋白检查和补碘检查；A multimodal colposcopy image acquisition module, configured to acquire a multimodal colposcopy image during a colposcopy examination acquired according to a first time sequence, wherein the colposcopy examination includes acetowhite examination, or acetowhite examination and iodine supplementation check;

拟诊模块，用于根据所有的所述多模态阴道镜图像的变化情况，确定拟诊结果，其中，所述拟诊结果包括：转化区病变大小、出现白上皮、醋白上皮表面和边界、血管、转化区类型、其它信息的性状，所述其他信息的性状包括可见持续性出血，可见易脆组织，可见白斑涂醋酸前存在是否可见；当所述阴道镜检查包括醋白检查和补碘检查时，所述拟诊结果还包括碘染色的性状；A diagnosis module, configured to determine the results of the diagnosis according to the changes of all the multimodal colposcopy images, wherein the results of the diagnosis include: the size of the lesion in the transformation zone, the appearance of white epithelium, the surface and boundary of the acetowhite epithelium , blood vessels, type of transformation zone, and other information traits, the other information traits include visible persistent bleeding, visible brittle tissue, visible vitiligo before application of acetic acid. During the iodine examination, the proposed diagnosis result also includes the characteristics of iodine staining;

评分模块，用于按照预先设定的评估信息，对所述拟诊结果进行评分，得到评估总分；The scoring module is used for scoring the proposed diagnosis result according to the preset evaluation information to obtain the total evaluation score;

转诊信息获取模块，用于获取转诊信息；其中，所述转诊信息包括：HPV检测信息、细胞学信息；A referral information acquisition module, used to acquire referral information; wherein, the referral information includes: HPV detection information and cytology information;

预测结果确定模块，用于根据所述转诊信息和所述评估总分，确定预测结果，其中所述预测结果包括病变类型和病变部位活检提示。The prediction result determination module is configured to determine the prediction result according to the referral information and the total evaluation score, wherein the prediction result includes the lesion type and the biopsy prompt of the lesion site.

本发明的发明目的三是提供一种阴道镜人工智能辅助质控管理系统，通过以下技术方案得以实现的：The third purpose of the invention of the present invention is to provide a colposcopy artificial intelligence-assisted quality control management system, which is achieved through the following technical solutions:

一种阴道镜人工智能辅助质控管理系统，包括存储器和处理器，所述存储器上存储有能够被处理器加载并执行如上方法的计算机程序。A colposcopy artificial intelligence-assisted quality control management system includes a memory and a processor, and the memory stores a computer program that can be loaded by the processor and execute the above method.

本发明的发明目的四是提供一种计算机可读存储介质，通过以下技术方案得以实现的：The fourth purpose of the present invention is to provide a computer-readable storage medium, which is achieved through the following technical solutions:

一种计算机可读存储介质，所述计算机可读存储介质中存储有指令，所述指令在计算机上执行时，使得所述计算机执行如上所述的方法。A computer-readable storage medium having instructions stored in the computer-readable storage medium, when the instructions are executed on a computer, cause the computer to execute the method as described above.

附图说明Description of drawings

图1为本申请实施例提供的一种阴道镜人工智能辅助质控管理方法的流程示意图；1 is a schematic flowchart of a colposcopy artificial intelligence-assisted quality control management method provided by the embodiment of the application;

图2为本申请实施例提供的一种图像采集质控的准备流程示意；2 is a schematic diagram of a preparation process for image acquisition quality control provided by an embodiment of the present application;

图3为本申请实施例提供的一种涂醋酸前评估的流程示意图；Fig. 3 is a kind of schematic flow sheet of evaluation before acetic acid coating provided by the embodiment of the application;

图4为本申请实施例的获取多模态阴道镜图像的具体流程示意图；4 is a schematic diagram of a specific flow of acquiring a multimodal colposcopy image according to an embodiment of the application;

图5为本申请实施例提供的一种阴道镜人工智能辅助质控管理装置的结构示意图；5 is a schematic structural diagram of a colposcopy artificial intelligence-assisted quality control management device provided in an embodiment of the application;

图6为本申请实施例提供的一种阴道镜人工智能辅助质控管理系统的结构示意图。FIG. 6 is a schematic structural diagram of a colposcopy artificial intelligence-assisted quality control management system according to an embodiment of the present application.

具体实施方式Detailed ways

以下结合附图1-附图6对本发明作进一步详细说明。The present invention will be described in further detail below with reference to accompanying drawings 1 to 6 .

本具体实施例仅是对本发明的解释，其并不是对本发明的限制，本领域技术人员在阅读完本说明书后可以根据需要对本实施例做出没有创造性贡献的修改，但只要在本发明的范围内都受到专利法的保护。This specific embodiment is only an explanation of the present invention, and it does not limit the present invention. Those skilled in the art can make modifications without creative contribution to the present embodiment as needed after reading this specification, but only within the scope of the present invention are protected by patent law.

为使本发明实施例的目的、技术方案和优点更加清楚，下面将结合本发明实施例中的附图，对本发明实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例是本发明一部分实施例，而不是全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有作出创造性劳动前提下所获得的所有其他实施例，都属于本发明保护的范围。In order to make the purposes, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments These are some embodiments of the present invention, but not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

另外，本文中术语“和/或”，仅是一种描述关联对象的关联关系，表示可以存在三种关系，例如，A和/或B，可以表示：单独存在A，同时存在A和B，单独存在B这三种情况。另外，本文中字符“/”，如无特殊说明，一般表示前后关联对象是一种“或”的关系。In addition, the term "and/or" in this document is only an association relationship to describe associated objects, indicating that there can be three kinds of relationships, for example, A and/or B, it can mean that A exists alone, A and B exist simultaneously, There are three cases of B alone. In addition, the character "/" in this text, unless otherwise specified, generally indicates that the related objects before and after are an "or" relationship.

相关技术的阴道镜检查方式是追踪阴道镜检查过程中阴道镜图像的特征间的关联变化直接进行诊断，仅针对特征的关联变化进行诊断容易造成诊断结果的不客观性，造成诊断的结果准确性低的问题。The colposcopy method of the related art is to track the correlation changes between the features of the colposcopy image during the process of colposcopy to make a diagnosis directly. Diagnosis only based on the correlation changes of the features will easily lead to the non-objectivity of the diagnosis results and the accuracy of the diagnosis results. low problem.

为解决上述技术问题，本实施例提供一种阴道镜人工智能辅助质控管理方法，阴道镜按照第一时序采集阴道镜检查过程中的多模态阴道镜图像，多模态表示利用多种颜色光线照射并获取图像，不同光线下的针对采集的重点不同，多模态阴道镜图像关联有时序以及多模态，能够更清楚的看到特征的细节以及特征的变化情况；根据多模态阴道镜图像确定拟诊结果，利用多模态特征以及关联对比提高了置信度，同时能够确定关注的病变位置；本实施例进一步按照预先设定的评估信息对拟诊结果得到评估总分，进而将转诊信息与评估总分结合进行预测，确定病变类型以及病变部位活检提示，更好的知道了拟诊，结果更加客观，提高了诊断的可靠性。In order to solve the above-mentioned technical problems, the present embodiment provides a colposcopy artificial intelligence-assisted quality control management method. The colposcope collects multimodal colposcopy images during the colposcopy process according to a first time sequence, and the multimodality means that multiple colors are used. Light irradiation and acquisition of images, the focus of acquisition under different light is different, multimodal colposcopy images are associated with timing and multimodality, which can more clearly see the details of features and changes in features; according to multimodal vaginal images Microscopic images are used to determine the results of the proposed diagnosis, and the multi-modal features and correlation comparison are used to improve the confidence, and at the same time, the location of the lesions of interest can be determined; this embodiment further obtains the total score for the evaluation of the results of the proposed diagnosis according to the pre-set evaluation information, and then the The referral information is combined with the total evaluation score to predict, determine the type of lesion and the biopsy prompt of the lesion site, better know the proposed diagnosis, the results are more objective, and the reliability of the diagnosis is improved.

综上可知，本实施例提供一种阴道镜人工智能辅助质控管理方法，请参考图1，图1为本申请实施例提供的一种阴道镜人工智能辅助质控管理方法的流程示意图，包括：To sum up, the present embodiment provides a colposcopy artificial intelligence-assisted quality control management method, please refer to FIG. 1, FIG. 1 is a schematic flowchart of a colposcopy artificial intelligence-assisted quality control management method provided by the embodiment of the application, including :

S110、获取按照第一时序采集的阴道镜检查的过程中的多模态阴道镜图像；S110, acquiring a multimodal colposcopy image in the process of colposcopy that is collected according to the first sequence;

其中，阴道镜检查包括醋白检查，或，醋白检查和补碘检查。Among them, colposcopy includes vinegar white inspection, or, vinegar white inspection and iodine supplementation inspection.

本实施例的执行主体为阴道镜人工智能辅助质控管理系统，可以为用户提供一个或者多个输入输出接口，输入输出接口与输入装置连接，该阴道镜人工智能辅助质控管理系统还通过设备端口与阴道镜系统连接，通过设备端口接收阴道镜系统采集的多模态阴道镜图像。其中，阴道镜系统包括阴道镜镜头、与阴道镜镜头连接的显示设备，阴道镜镜头包括光源、用于对待检测部位进行多模态阴道镜图像采集的图像采集设备，待检测部位可以是整个宫颈，还可以是宫颈的目标区域，本实施例不再进行限定；显示设备用于显示图像采集设备采集的多模态阴道镜图像。进一步的，显示设备当然还可以显示相关的多模态阴道镜图像显示的比例尺、操作时间和相关检查信息。同时，显示设备还可以利用预设接口与阴道镜人工智能辅助质控管理系统连接，进行多模态阴道镜图像的数据传输。The executive body of this embodiment is a colposcopy artificial intelligence-assisted quality control management system, which can provide users with one or more input and output interfaces, and the input and output interfaces are connected to the input device. The colposcopy artificial intelligence-assisted quality control management system also uses equipment The port is connected with the colposcopy system, and the multimodal colposcopy image collected by the colposcopy system is received through the device port. The colposcopy system includes a colposcopy lens and a display device connected to the colposcopy lens. The colposcopy lens includes a light source and an image acquisition device for collecting multi-modal colposcopy images of the site to be tested. The site to be tested may be the entire cervix. , which can also be the target area of the cervix, which is not limited in this embodiment; the display device is used to display the multimodal colposcopy image collected by the image collection device. Further, of course, the display device can also display the scale, operation time and relevant examination information of the multimodal colposcopy image display. At the same time, the display device can also use a preset interface to connect with the colposcopy artificial intelligence-assisted quality control management system to perform data transmission of multimodal colposcopy images.

由于宫颈的生理结构和上皮细胞组织生物学特点，需要利用照明光线照射宫颈组织，实现对宫颈组织的细节的鉴别，该照明光线可以是白光、红光、绿光、蓝光中的任意一种或者多种，不同颜色光照下采集到的图像的细节特征不同，例如，白光下可以采集到宫颈的常规状态，而绿光下能够采集到血管的细节图像；同时，对于阴道镜检查，阴道镜系统还要按照阴道镜人工智能辅助质控管理系统设定的第一时序进行检查，最终在多颜色光灯、多时间段采集到多维度的阴道镜图像，作为多模态阴道镜图像，可以理解的是，多模态阴道镜图像之间的关联变化，提供了阴道镜诊断的基础。阴道镜检查的项目包括醋白检查，或醋白检查和补碘检查，第一时序包括阴道镜检查过程设定多个采集图像时刻、设定情形下的照明光线选择以及采集图像；在一种可实现的实施方式中，第一时序为醋白检查对应的时序，在另一种可实现的实施方式中，第一时序为醋白检查和补碘检查对应的时序。Due to the physiological structure of the cervix and the biological characteristics of epithelial cells, it is necessary to use illumination light to illuminate the cervical tissue to realize the identification of the details of the cervical tissue. The illumination light can be any one of white light, red light, green light, blue light or There are many kinds, the details of the images collected under different colors of light are different, for example, the general state of the cervix can be collected under white light, while the detailed images of blood vessels can be collected under green light; at the same time, for colposcopy, the colposcopy system It is also necessary to check according to the first sequence set by the colposcopy artificial intelligence-assisted quality control management system, and finally collect multi-dimensional colposcopy images in multi-color lights and multiple time periods. As a multi-modal colposcopy image, it is understandable Interestingly, the correlation changes between multimodal colposcopy images provide the basis for colposcopy diagnosis. The items of colposcopy include vinegar white inspection, or vinegar white inspection and iodine supplementation inspection, and the first sequence includes setting a plurality of time for collecting images during colposcopy, selection of illumination light under the setting situation, and collecting images; In an achievable embodiment, the first sequence is the sequence corresponding to the vinegar white inspection, and in another achievable embodiment, the first sequence is the sequence corresponding to the vinegar white inspection and the iodine supplementation inspection.

以阴道镜检查仅包括醋白检查为例，本实施例提供一种具体的醋白检查的检测与图像采集过程。醋白检查为利用5％冰醋酸通过涂抹宫颈组织后在特定时间内观察组织上皮的特殊变化来辅助诊断；宫颈组织的上皮细胞在5％冰醋酸作用下，改变了光线的作用特性，使宫颈成像得到的阴道镜图像在涂抹前后在有限的时间内观察，有显著的形态区别，正常的宫颈组织的上皮细胞涂抹后会先有醋酸白色上皮出现，在有限时间内醋酸白色上皮消失；存在病变的宫颈组织会有醋酸白色上皮，且醋酸白色上皮在有限时间内不能消失。用大棉签采用喷淋的方法涂抹10s后开始观察，边观察边涂抹，60s后结束涂抹，发现醋酸白色上皮的最快形成区域位置；并按照第一时序具体可以是每隔30s采集一张阴道镜图像，150s结束；也就是说在醋酸反应30s后、60s、90s、120s、150s分别采集阴道镜图像，同时，在固定间隔采集阴道镜图像的同时，还需要按照第一时序中的光源设定规则采集不同颜色光源(如白色光源、绿色光源、红色光源)下的阴道镜图像，最终得到一系列的多模态阴道镜图像。针对检查过程中醋白上皮变化的重点关注：形成时间、增厚时间、消退时间及范围(占几象限)；是否出现异常血管；比对与涂醋酸前原始图像相同位置是否血运丰富(图像显血管红)。根据醋白上皮重点关注信息的权重，用“+”提示重点关注部位置信度信息(依据重点关注信息的权重)。Taking the colposcopy examination only including the acetowhite examination as an example, this embodiment provides a specific detection and image acquisition process of the acetowhite examination. The acetowhite examination is to use 5% glacial acetic acid to assist the diagnosis by observing the special changes of the tissue epithelium within a specific time after smearing the cervical tissue; the epithelial cells of the cervical tissue change the action characteristics of light under the action of 5% glacial acetic acid, making the cervical The colposcopy images obtained by imaging are observed in a limited time before and after smearing, and there are significant morphological differences. After smearing, the acetowhite epithelium of normal cervical tissue will appear first, and the acetowhite epithelium will disappear within a limited time; there are lesions. The cervical tissue will have acetowhite epithelium, and the acetowhite epithelium cannot disappear for a limited time. Use a large cotton swab to spray for 10s and start observation, observe while smearing, and finish smearing after 60s, and find the location of the fastest-forming area of acetowhite epithelium; and according to the first sequence, a vagina can be collected every 30s. The endoscopy image ends in 150s; that is to say, after the acetic acid reaction for 30s, 60s, 90s, 120s, and 150s, the colposcopy images are collected respectively. The colposcopy images under different color light sources (such as white light source, green light source, red light source) are collected according to certain rules, and finally a series of multimodal colposcopy images are obtained. Focus on the changes of the acetowhite epithelium during the examination: the formation time, thickening time, regression time and range (occupying several quadrants); whether there are abnormal blood vessels; vascular red). According to the weight of the focus information of the acetowhite epithelium, "+" is used to indicate the location reliability information of the focus part (according to the weight of the focus information).

综上可知，本实施例在图像采集的环节，能够按照检查规范流程按照第一时序进行图像采集，提高了规范性，不仅有经验的医护人员可以操作，经验不足的医护人员也可以操作，不依赖于个人经验，能够提升基层诊疗水平。From the above, it can be seen that in the image acquisition process of this embodiment, the image acquisition can be performed in accordance with the first time sequence according to the inspection standard process, which improves the standardization. Not only experienced medical staff can operate, but also inexperienced medical staff can operate. Relying on personal experience can improve the level of primary care.

进一步的，为了能够清楚的得到阴道镜图像的细节特征，在阴道镜人工智能辅助质控管理系统控制阴道镜系统进行阴道镜图像采集过程中，阴道镜人工智能辅助质控管理系统接收到当前时刻的阴道镜图像后，在确定该阴道镜图像中存在异常情况，还可以自动确定异常情况对应的局部区域，控制阴道镜系统进行局部特征细节再采集。Further, in order to clearly obtain the detailed characteristics of the colposcopy image, during the colposcopy AI-assisted quality control management system controls the colposcopy system to collect the colposcopy image, the colposcopy AI-assisted quality control management system receives the current time. After the colposcopy image is obtained, when it is determined that there is an abnormal situation in the colposcopy image, the local area corresponding to the abnormal situation can also be automatically determined, and the colposcopy system can be controlled to collect local feature details again.

为了能够较好的弥补醋白检测识别的假阳性和漏诊，完善对SPI(宫颈HPV亚临床感染)/HPV(人乳头瘤病毒)感染程度和部位的鉴别，可以进行补碘检查。对应的，阴道镜检查包括醋白检查和补碘检查，补碘检查为利用5％卢戈氏碘溶液染色后采集阴道镜图像，可以在涂抹5％卢戈氏碘溶液30秒后开始控制阴道镜系统按照第一时序中对应补碘检查的子时序采集阴道镜图像，可以是每隔30s采集一张阴道镜图像，150s结束；也就是说在涂抹5％卢戈氏碘溶液30s后、60s、90s、120s、150s分别采集阴道镜图像，采集目的是发现上皮碘吸收染色状态，涂醋酸后按时序变化图像重点关注部位及碘染后的相同位置图像是否染色进行关联对比，可增强对化生上皮的识别率。In order to better make up for the false positives and missed diagnoses identified by acetowhite detection, and to improve the identification of the degree and location of SPI (cervical HPV subclinical infection)/HPV (human papillomavirus) infection, iodine supplementation can be performed. Correspondingly, colposcopy includes vinegar white examination and iodine supplementation examination. The iodine supplementation examination is to collect colposcopy images after staining with 5% Lugol's iodine solution, and the vagina can be controlled after 30 seconds of application of 5% Lugol's iodine solution. The endoscopy system collects colposcopy images according to the sub-sequence corresponding to the iodine supplementation examination in the first sequence, which can be a colposcopy image every 30s, and ends in 150s; that is, after applying 5% Lugol's iodine solution for 30s, 60s Colposcopy images were collected at 90s, 120s, and 150s, respectively. The purpose of the collection was to find out the state of iodine absorption and staining in the epithelium. After applying acetic acid, the focus of the image was changed according to the time series, and the images at the same position after iodine staining were correlated to compare whether they were stained. Identification of raw epithelium.

S120、根据所有的多模态阴道镜图像的变化情况，确定拟诊结果；S120, according to the changes of all the multimodal colposcopy images, determine the results of the suspected diagnosis;

其中，拟诊结果包括：转化区病变大小、出现白上皮、醋白上皮表面和边界、血管、转化区类型、其它信息的性状，其他信息的性状包括可见持续性出血，可见易脆组织，可见白斑涂醋酸前存在是否可见；当阴道镜检查包括醋白检查和补碘检查时，拟诊结果还包括碘染色的性状；Among them, the proposed diagnosis results include: the size of the lesion in the transformation zone, the appearance of white epithelium, the surface and border of the acetowhite epithelium, blood vessels, the type of the transformation zone, and other informative traits. Other informative traits include visible persistent bleeding, visible brittle tissue, visible Whether the vitiligo is visible before applying acetic acid; when colposcopy includes acetowhite inspection and iodine supplementation inspection, the proposed diagnosis also includes iodine staining;

本实施例是基于多模态阴道镜图像的变化情况确定拟诊结果，确定的手段可以是医护人员确定，还可以是采用人工智能的手段确定；当然，当医护人员为经验丰富的医师时，得到拟诊结果，并将对应的多模态阴道镜图像进行标记，得到标记图像，将标记图像和拟诊结果存储至数据库，作为人工智能训练的后续数据，以便能够随着训练数据的丰富逐渐提高人工智能的检测可靠性。In this embodiment, the simulated diagnosis result is determined based on the change of the multimodal colposcopy image, and the determination method may be determined by medical staff or artificial intelligence; of course, when the medical staff is an experienced doctor, Obtain the simulated diagnosis results, mark the corresponding multi-modal colposcopy images, obtain the labeled images, and store the labeled images and the simulated diagnosis results in the database as the follow-up data for artificial intelligence training, so as to be able to gradually increase with the enrichment of training data. Improve the detection reliability of artificial intelligence.

其中，转化区病变大小的性状包括：尺寸或所占象限，具体包括第一种性状小于5mm；第二种性状5～15mm或占2个象限；第三种性状大于15mm或占3～4个象限，大小不确定。出现白上皮的性状包括：第一种性状无或透明；第二种性状薄雾状、半透明乳白色；第三种性状明显、不透明白色。醋白上皮表面和边界的性状包括：第一种性状弥漫；第二种性状边界清晰但不规则，锯齿状、“地图样”，卫星样病灶；第三种性状，边界清晰，表面不平，包含袖口样腺开口。血管的性状包括：第一种性状细小，均匀；第二种性状规则；第三种性状粗大点状或镶嵌，或异型血管。转化区类型的性状：第一种性状Ⅰ型转化区病变、第二种性状Ⅱ型转化区病变、第三种性状Ⅲ型转化区病变。其它信息的性状，其中，其他信息的性状包括可见持续性出血，可见易脆组织，可见白斑涂醋酸前存在是否可见。碘染色的性状：第一种性状着色显棕色；第二种性状部分着色微弱或斑点状黄；第三种性状不着色明显的黄。Among them, the traits of the size of the lesions in the transformation zone include: size or quadrant occupied, specifically including that the first trait is less than 5 mm; the second trait is 5 to 15 mm or occupies 2 quadrants; the third trait is greater than 15 mm or occupies 3 to 4 Quadrant, indeterminate size. The traits of white epithelium include: the first trait is absent or transparent; the second trait is hazy, translucent milky white; the third trait is obvious, opaque white. Traits of acetowhite epithelial surfaces and borders include: the first trait, diffuse; the second trait, well-defined but irregularly bordered, jagged, "geographical", satellite-like lesions; and the third trait, well-defined, uneven surface, containing Cuff-like gland openings. The traits of blood vessels include: the first trait is small and uniform; the second trait is regular; the third trait is thick dotted or mosaic, or atypical blood vessels. Traits of transformation zone types: the first trait type I transformation zone lesions, the second trait type II transformation zone lesions, and the third trait type III transformation zone lesions. Other informative traits include visible persistent bleeding, visible brittle tissue, visible vitiligo prior to acetic acid application or not. Characters of iodine dyeing: the first character is colored brown; the second character is partially colored faint or spotted yellow; the third character is not clearly colored yellow.

S130、按照预先设定的评估信息，对拟诊结果进行评分，得到评估总分；S130, according to the preset evaluation information, score the results of the proposed diagnosis to obtain a total evaluation score;

本实施例提供一种预先设定的评估信息，请参考表1，当然，还能够采用其他的评估信息，用户可自定义设置。根据表1中的评分标准确定评估总分。例如，当转化区病变大小为5～15mm或占2个象限；出现白上皮为明显、不透明白色；醋白上皮表面和边界为弥漫；血管为粗大点状或镶嵌；碘染色的性状为部分着色微弱或斑点状黄，则得到的评估总分为1+2+0+2+1＝6。This embodiment provides a preset evaluation information, please refer to Table 1, of course, other evaluation information can also be used, and the user can customize the settings. The total score for the assessment was determined according to the scoring criteria in Table 1. For example, when the size of the lesion in the transformation zone is 5-15 mm or occupies 2 quadrants; the white epithelium is obvious and opaque white; the surface and border of the acetowhite epithelium is diffuse; the blood vessels are thick punctate or mosaic; Faint or speckled yellow, an overall score of 1+2+0+2+1=6 is obtained.

表1为预先设定的评估信息Table 1 is the pre-set evaluation information

S140、获取转诊信息；S140. Obtain referral information;

其中，转诊信息包括：HPV检测信息、细胞学信息；The referral information includes: HPV testing information, cytology information;

HPV检测信息包括阴性或非16/18型、HPV16/18型。细胞学信息包括细胞学正常、≥ASC-H(不典型鳞状上皮细胞，倾向于高度病变)。HPV test information includes negative or non-16/18 type, HPV16/18 type. Cytological information included normal cytology, ≥ASC-H (atypical squamous cells, tending to be highly diseased).

当然，转诊信息还可以包括年龄，用户可根据实际需求进行设置，只要是能够实现本实施例的目的即可。Of course, the referral information may also include age, and the user may set it according to actual needs, as long as the purpose of this embodiment can be achieved.

S150、根据转诊信息和评估总分，确定预测结果；S150. Determine the prediction result according to the referral information and the total evaluation score;

其中预测结果包括病变类型和病变部位活检提示。The predicted results included the type of lesion and the biopsy tip of the lesion site.

通过将转诊信息关联评估总分，能够更好地指导拟诊，结果更客观。By correlating the referral information with the total score of the assessment, the proposed diagnosis can be better guided, and the results are more objective.

其中，将评估总分结合转诊信息确定预测结果，其中，预测结果包括病变类型以及病变部位的提示。在阴道镜检查时一般存在的病变类型包括NILM(未见上皮内细胞病变)、LSIL(Low-grade Squamous Intraepithelial Lesion,低度鳞状上皮内病变)、HSIL(high-grade squamous intraepithelial lesion,高分化鳞状上皮内病变)、CA(癌)；病变部位活检提示包括建议活检、建议多点活检。在一种可实现的实施方式中，确定预测结果的方式可以是根据系统内存储的预设转诊信息、评估总分以及预测结果的对应关系中确定与当前的转诊信息、当前的评估总分对应的当前的预测结果；可以理解的是，该对应关系是根据有经验的医疗工作者在实际工作中积累得到的。在另一种可实现的实施方式中，确定预测结果的方式包括：根据评估总分确定初始预测结果；根据转诊信息从初始预测结果进行筛选，得到预测结果。其中，病变部位包括但是不限定于：具有评分的醋白上皮表面和边界、血管、碘染色处。Among them, the total score of the evaluation is combined with the referral information to determine the prediction result, wherein the prediction result includes the lesion type and the hint of the lesion location. The types of lesions commonly present during colposcopy include NILM (no intraepithelial lesion seen), LSIL (Low-grade Squamous Intraepithelial Lesion, low-grade squamous intraepithelial lesion), HSIL (high-grade squamous intraepithelial lesion, well-differentiated) Squamous intraepithelial lesions), CA (carcinoma); biopsy tips for lesions include recommended biopsy and recommended multi-point biopsy. In an achievable implementation manner, the method of determining the prediction result may be to determine the current referral information, the current evaluation total score and the corresponding relationship between the preset referral information, the evaluation total score and the prediction result stored in the system. The corresponding current prediction results; it can be understood that the corresponding relationship is accumulated according to the actual work of experienced medical workers. In another achievable implementation manner, the method of determining the prediction result includes: determining the initial prediction result according to the total evaluation score; and screening the initial prediction result according to the referral information to obtain the prediction result. Wherein, the lesions include but are not limited to: acetowhite epithelial surfaces and borders with scores, blood vessels, and iodine staining.

基于上述技术方案，本实施例按照第一时序采集阴道镜检查过程中的多模态阴道镜图像，保证采集的质控管理，即使是经验不足的医疗工作者也可以执行，然后，根据获取的时序变化的多模态阴道镜图像的关联特征共同确定拟诊结果，根据该拟诊结果能够找出重点关注的信息；基于人工智能，利用预先设定的评估信息对拟诊结果进行评分，得到评估总分；将该评估总分关联转诊信息共同对当前的阴道镜诊断进行预测，让评估更加客观，提高了预测结果的可靠性。Based on the above technical solution, the present embodiment collects the multimodal colposcopy images during the colposcopy process according to the first time sequence, so as to ensure the quality control management of the collection, which can be performed even by inexperienced medical workers. The correlation characteristics of the multimodal colposcopy images of the time series change jointly determine the suspected diagnosis result, and the key information can be found according to the suspected diagnosis result; Total evaluation score; the evaluation total score is associated with referral information to predict the current colposcopy diagnosis, which makes the evaluation more objective and improves the reliability of the prediction results.

进而，为了能够提高阴道镜管理的质控，尽可能减少图像采集不清楚或者不完善的问题，本实施例获取按照第一时序采集的阴道镜检查的过程中的多模态阴道镜图像之前，请参考图2，图2为本申请实施例提供的一种图像采集质控的准备流程示意，包括：Furthermore, in order to improve the quality control of colposcopy management and minimize the problem of unclear or incomplete image acquisition, in this embodiment, before acquiring the multimodal colposcopy image in the process of colposcopy acquired according to the first time sequence, Please refer to FIG. 2. FIG. 2 is a schematic diagram of a preparation process for image acquisition quality control provided by an embodiment of the present application, including:

S11、获取镜头参数下的阴道镜图像的清晰度，镜头参数包括镜头位置，镜头的各光源的照度、均匀值以及色温；S11. Acquire the clarity of the colposcope image under the lens parameters, the lens parameters include the lens position, the illuminance, uniform value and color temperature of each light source of the lens;

本实施例中的阴道镜图像是没有涂覆液体的宫颈图像。本实施例中阴道镜镜头可以包括多个光源，每个光源对应的参数包括照度、均匀值和色温。光源可以包括绿光光源、红光光源、白光光源，每个光源要在合适的照度、均匀值以及色温下得到效果最好的阴道镜图像。The colposcopy image in this example is an image of the cervix without the coating liquid. In this embodiment, the colposcope lens may include multiple light sources, and parameters corresponding to each light source include illuminance, uniform value, and color temperature. The light sources can include green light sources, red light sources, and white light sources, and each light source should obtain the best colposcopy image under appropriate illumination, uniform value and color temperature.

S12、发送镜头参数调整指令至阴道镜镜头，以使阴道镜镜头调整镜头参数，实现各光源下的阴道镜图像的清晰度至设定清晰度；S12, sending a lens parameter adjustment instruction to the colposcope lens, so that the colposcope lens adjusts the lens parameters, so that the clarity of the colposcope image under each light source can reach the set clarity;

本实施例的目的是调整镜头参数，使阴道镜图像在各光源下的清晰度均为设定清晰度，完成对多光源下的图像的清晰度的校准，以便后续采集到的多模态阴道镜图像均为高清图像。The purpose of this embodiment is to adjust the lens parameters, so that the clarity of the colposcopy image under each light source is the set definition, and to complete the calibration of the clarity of the image under multiple light sources, so that the multi-modal vagina collected later can be obtained. The mirror images are high-definition images.

本实施例能够根据阴道镜图像的清晰度和设定清晰度的大小自动调整阴道镜镜头的图像采集设备位置，实现对图像采集设备的工作距离的自动调整，以便阴道镜图像的清晰度达到设定清晰度，该设定清晰度保证能够清楚看到阴道镜图像的细节。可见，本步骤是在成像环节，根据实际的成像效果来调整镜头工作参数，以便得到清晰的图像。This embodiment can automatically adjust the position of the image acquisition device of the colposcope lens according to the clarity of the colposcope image and the size of the set clarity, so as to realize the automatic adjustment of the working distance of the image acquisition equipment, so that the clarity of the colposcope image can reach the set point. The set definition ensures that the details of the colposcopy image can be clearly seen. It can be seen that this step is to adjust the working parameters of the lens according to the actual imaging effect in the imaging process, so as to obtain a clear image.

进一步的，还可以在镜头参数调整的过程中进行调整提示，利用微型激光测距功能，提示当前工作距离和景深，便于医护人员了解当前的镜头参数调整情况。Further, adjustment prompts can also be made during the adjustment of lens parameters, and the micro-laser ranging function can be used to prompt the current working distance and depth of field, so that medical staff can understand the current adjustment of lens parameters.

S13、确定设定清晰度的阴道镜图像是否存在异物；S13. Determine whether there is a foreign body in the colposcopy image of the set definition;

阴道镜镜头调整镜头参数，实现各光源下的阴道镜图像的清晰度至设定清晰度后，确定阴道镜图像是否存在异物，该异物非宫颈组织，可能是由于棉签带的棉絮或者其他外来污染物。若是，执行步骤S14，若否，执行步骤S15。After adjusting the lens parameters of the colposcope lens to achieve the clarity of the colposcopy image under each light source to the set clarity, determine whether there is a foreign body in the colposcopy image. The foreign body is not cervical tissue, which may be due to cotton wool or other foreign pollution thing. If yes, go to step S14, if no, go to step S15.

S14、发送提示信息至显示设备，提示信息用于提示医护人员去除异物；S14, sending prompt information to the display device, the prompt information is used to prompt the medical staff to remove the foreign body;

S15、按照校准后的镜头参数控制阴道镜镜头进行多模态阴道图像采集。S15 , controlling the colposcope lens according to the calibrated lens parameters to collect multimodal vaginal images.

若存在异物，则发送提示信息至显示设备上，以便医护人员看到后去除异物，使宫颈组织暴露出来，保证了后续采集的多模态阴道镜图像清楚且无异物。If there is a foreign body, a prompt message will be sent to the display device, so that the medical staff can remove the foreign body and expose the cervical tissue to ensure that the multimodal colposcopy image collected subsequently is clear and free of foreign bodies.

通过上述的阴道镜初始检测的质量控制，尽可能减少图像采集不清楚或者有异物的客观问题，提高了后续检测的准确性。Through the quality control of the above-mentioned initial detection of colposcopy, objective problems of unclear image acquisition or foreign bodies are minimized, and the accuracy of subsequent detection is improved.

进一步的，为了使医护人员能够清楚看到宫颈组织的阴道镜图像，还可以设定放大倍数，以使显示设备能够显示设定倍数的阴道镜图像，以便医护人员观察。Further, in order to enable the medical staff to clearly see the colposcopy image of the cervical tissue, the magnification can also be set, so that the display device can display the colposcopy image of the set magnification for the medical staff to observe.

基于上述实施例，在完成对成像的质控后，为了提高检查的效率，阴道镜人工智能辅助质控管理方法还包括涂醋酸前评估的过程，请参考图3，图3为本申请实施例提供的一种涂醋酸前评估的流程示意图，包括：Based on the above embodiment, after completing the quality control of imaging, in order to improve the efficiency of inspection, the artificial intelligence-assisted quality control management method for colposcopy also includes a process of evaluating before applying acetic acid, please refer to FIG. 3 , which is an embodiment of the application Provided is a schematic flow chart of assessment before applying acetic acid, including:

S16、当宫颈部位涂覆生理盐水后，获取按照第二时序采集的多张生理盐水涂覆后的生理盐水阴道镜图像；S16, after the cervix is coated with the physiological saline, obtain a plurality of saline colposcopy images after the saline is coated and collected according to the second time sequence;

第二时序包括按照涂覆后的30s后，间隔设定时间采集生理盐水阴道镜图像。The second time sequence includes collecting normal saline colposcopy images at set time intervals after 30 s after coating.

S17、对所有生理盐水阴道镜图像进行特征识别，确定是否具有可疑癌特征；S17. Perform feature identification on all saline colposcopy images to determine whether they have suspicious cancer features;

可疑癌特征包括：肿瘤或新生肿物、外生型病变、脆性血管、表面不规则、坏死、溃疡。Suspected features of cancer include: neoplastic or neoplastic mass, exophytic lesions, fragile vessels, surface irregularities, necrosis, ulceration.

本实施例是在阴道镜人工智能辅助质控管理系统的本地AI算法中动态的特征识别，确认是否具有可疑癌特征。进一步的，同时还可以存储所有的生理盐水阴道镜图像以便有经验的医疗工作者进行确认，得到确认结果，将该生理盐水阴道镜图像和确认结果作为训练样本，当训练样本达到设定数量时进行训练，提高本地AI算法的精准度。In this embodiment, dynamic feature identification is performed in the local AI algorithm of the colposcopy artificial intelligence-assisted quality control management system to confirm whether there is a suspicious cancer feature. Further, at the same time, all the normal saline colposcopy images can also be stored for confirmation by experienced medical workers to obtain confirmation results, and the normal saline colposcopy images and confirmation results can be used as training samples, when the training samples reach the set number. Conduct training to improve the accuracy of local AI algorithms.

若具有，则提示多点活检。If present, multiple biopsy is suggested.

S18、若不具有，则在显示界面显示待评估信息；S18. If not, display the information to be evaluated on the display interface;

待评估信息包括：宫颈上皮雌激素作用影响、绝经后上皮萎缩性改变，检查前是否做过雌激素治疗、妊娠新改变存在、阴道镜下确定急性感染存在、镜下能充分观察到整个宫颈和阴道上段、确认是否接受过治疗或镜下可见治疗后改变、上皮纤维化或镜下不同程度可见。The information to be evaluated includes: the effect of estrogen on the cervical epithelium, postmenopausal epithelial atrophic changes, whether estrogen therapy has been done before the examination, the presence of new changes in pregnancy, the presence of acute infection confirmed by colposcopy, the ability to fully observe the entire cervix and Upper vaginal segment, confirm whether it has received treatment or microscopically visible changes after treatment, epithelial fibrosis or microscopically visible to varying degrees.

待评估信息用于医护人员进行选择，待评估信息包括检查7项。其中，宫颈上皮雌激素作用影响的评估选项包括：好、差；绝经后上皮萎缩性改变，检查前是否做过雌激素治疗的评估选项包括：有、无；妊娠新改变存在的评估选项包括：有、无；阴道镜下确定急性感染存在(病毒、真菌、细菌的评估选项包括：有、无；镜下能充分观察到整个宫颈和阴道上段的评估选项包括：有、无；确认是否接受过治疗或镜下可见治疗后改变的评估选项包括：有、无；上皮纤维化或镜下不同程度可见的评估选项包括：有、无。The to-be-evaluated information is used for selection by medical staff, and the to-be-evaluated information includes 7 inspection items. Among them, the evaluation options for the effect of estrogen on the cervical epithelium include: good, poor; postmenopausal epithelial atrophy changes, and the evaluation options for whether estrogen therapy has been done before the examination include: yes or no; the assessment options for the presence of new changes in pregnancy include: Yes, no; colposcopy to determine the presence of acute infection (virus, fungus, bacteria assessment options include: yes, no; microscopic assessment options that can fully observe the entire cervix and upper vaginal segment include: yes, no; confirm whether received Evaluation options for treatment or microscopically visible changes after treatment include: yes, no; evaluation options for epithelial fibrosis or microscopically visible changes to varying degrees include: yes, no.

S19、接收到当前医疗工作者从待评估信息中选择的已评估信息；S19. Receive the evaluated information selected by the current medical worker from the information to be evaluated;

根据该已评估信息确定是否可执行阴道镜检查。Based on this assessed information, it is determined whether colposcopy can be performed.

S20、根据已评估信息确定阴道镜检查条件是否充分，若是，则执行获取按照第一时序采集的阴道镜检查的过程中的多模态阴道镜图像的步骤。S20. Determine whether the colposcopy conditions are sufficient according to the evaluated information, and if so, execute the step of acquiring a multimodal colposcopy image during the colposcopy process acquired according to the first time sequence.

若不充分，则需要治疗/处理后再检查，或补充其他检查。If it is not enough, it needs to be checked after treatment/treatment, or supplemented with other checks.

通过上述技术方案可知，本实施例能够对进行醋酸前评估，以避免无效的检查。As can be seen from the above technical solutions, this embodiment can perform pre-acetic acid evaluation to avoid invalid inspection.

进一步的，为了由于存在漏检情况，会影响拟诊结果的确定，请参考图4，图4为本申请实施例的获取多模态阴道镜图像的具体流程示意图，包括：Further, in order to affect the determination of the suspected diagnosis result due to the existence of missed detection, please refer to FIG. 4 , which is a schematic diagram of a specific flow of obtaining a multimodal colposcopy image according to the embodiment of the application, including:

S111、获取按照第一时序采集的阴道镜检查的过程中的多模态阴道镜图像；S111, acquiring a multimodal colposcopy image during the colposcopy process collected according to the first sequence;

该多模态阴道镜图像的获取之后会立刻上传至阴道镜人工智能辅助质控管理系统。The multimodal colposcopy image will be uploaded to the colposcopy AI-assisted quality control management system immediately after the acquisition.

S112、判断当前时刻是否存在多模态阴道镜图像的漏检情况；S112, judging whether there is a missed detection of the multimodal colposcopy image at the current moment;

若存在，则执行步骤S113，若不存在，则执行步骤S115。If there is, go to step S113, if not, go to step S115.

S113、进行漏检提示，并控制阴道镜镜头进行再次采集，得到补充多模态阴道镜图像；S114、根据漏检的图像替换成补充多模态阴道镜图像；S113, performing a missed detection prompt, and controlling the colposcopy lens to collect again, to obtain a supplementary multimodal colposcopy image; S114, replacing the missed detection image with a supplementary multimodal colposcopy image;

S115、确定不存在漏检情况。S115, it is determined that there is no missing detection.

该漏检情况可以表示为该漏检多模态阴道镜图像的概率或者程度，还可以包括获取多模态阴道镜图像的时刻、检测部门、检测人员等信息，以便作为参考信息进行溯源。The missed detection situation may be expressed as the probability or degree of the missed detection of the multimodal colposcopy image, and may also include information such as the time when the multimodal colposcopy image was obtained, the detection department, and the detection personnel, so as to be used as reference information for traceability.

判断漏检情况的方式可以是根据多模态阴道镜图像的图像质量(在实际操作过程中，由于被检测者的运动造成的清晰度低)、多模态阴道镜图像与对应光源匹配程度。The way of judging the missed detection can be based on the image quality of the multimodal colposcopy image (in the actual operation process, the resolution is low due to the movement of the detected person), and the matching degree between the multimodal colposcopy image and the corresponding light source.

由此，在一种可实现的实施例方式中，判断当前时刻是否存在多模态阴道镜图像的漏检情况，包括：确定当前时刻的多模态阴道镜图像的图像质量是否达到标准质量；若是，则确定没有漏检情况；若否，则确定出现漏检情况。具体的，确定多模态阴道镜图像的图像质量的方式包括：通过确定多模态阴道镜图像的亮度以及曝光比确定图像质量。Thus, in an achievable embodiment, judging whether there is a missed detection of the multimodal colposcopy image at the current moment includes: determining whether the image quality of the multimodal colposcopy image at the current moment reaches a standard quality; If so, it is determined that there is no missed detection; if not, it is determined that there is a missed detection. Specifically, the method of determining the image quality of the multimodal colposcopy image includes: determining the image quality by determining the brightness and exposure ratio of the multimodal colposcopy image.

在另一种可实现的实施例方式中，判断当前时刻是否存在多模态阴道镜图像的漏检情况，包括：确定当前多模态阴道镜图像对应的光源颜色是否对应指定光源颜色，若是，则确定没有漏检情况；若否，则确定出现漏检情况。In another achievable embodiment, judging whether there is a missed detection of the multimodal colposcopy image at the current moment includes: determining whether the color of the light source corresponding to the current multimodal colposcopy image corresponds to the color of the specified light source, and if so, Then it is determined that there is no missed detection; if not, it is determined that there is a missed detection.

其中，针对进行漏检提示做进一步阐述，可以是以语音的形式播放漏检提示，其中，该漏检提示包括但是不限定于漏检提醒信息、指定颜色光源、检测时间；还可以是发送漏检提示对应的漏检信息至客户端设备；还可以控制在显示设备显示与漏检提示对应的警示信息，以便规范用户的内窥镜检查操作。Among them, to further elaborate on the missed detection prompt, the missed detection prompt can be played in the form of voice, wherein the missed detection prompt includes but is not limited to missed detection reminder information, specified color light source, detection time; The missed detection information corresponding to the detection prompt is sent to the client device; the warning information corresponding to the missed detection prompt can also be controlled to be displayed on the display device, so as to standardize the user's endoscopy operation.

进一步的，本实施例提供两种确定拟诊结果的方式，具体的，根据所有的多模态阴道镜图像的变化情况，确定拟诊结果，包括：Further, this embodiment provides two ways to determine the results of the simulated diagnosis. Specifically, according to the changes of all multimodal colposcopy images, the results of the simulated diagnosis are determined, including:

在显示界面按照第一时序显示所有的多模态阴道镜图像，以便当前医疗工作者根据所有的多模态阴道镜图像的变化情况输入拟诊结果；All the multimodal colposcopy images are displayed on the display interface according to the first sequence, so that the current medical worker can input the simulated diagnosis results according to the changes of all the multimodal colposcopy images;

获取用户确定的拟诊结果；Obtain the proposed diagnosis result determined by the user;

或，or,

按照第一时序发送所有的多模态阴道镜图像至云端服务器，以便云端服务器对多模态阴道镜图像进行自动特征识别，并根据特征的变化情况，生成拟诊结果；Send all the multimodal colposcopy images to the cloud server according to the first time sequence, so that the cloud server can automatically identify the features of the multimodal colposcopy images, and generate simulated diagnosis results according to the changes of the characteristics;

获取云端服务器发送的拟诊结果。Obtain the suspected diagnosis result sent by the cloud server.

本实施例提供两种方式，一种是用户输入，另一种是云端服务器智能识别。This embodiment provides two methods, one is user input, and the other is intelligent recognition by the cloud server.

对于用户输入的方式，可以是在显示第一时序的多模态阴道镜图像后，医疗工作者根据自身的知识储备确定拟诊结果，该拟诊结果可以在屏幕上显示各种选择项目。例如，转化区病变大小的选择项目包括小于5mm，5-15mm或占2个象限，大于15mm或占3-4个象限大小不确定，用户可根据实际情况进行勾选。出现白上皮的选择项目包括无或透明，薄雾状、半透明乳白色，明显、不透明白色，用户可根据实际情况进行勾选。醋白上皮表面和边界的选择项目包括：弥漫；边界清晰但不规则，锯齿状、“地图样”，卫星样病灶；边界清晰，表面不平，包含袖口样腺开口，用户可根据实际情况进行勾选。最终，根据用户勾选的所有内容确定拟诊结果。For the user input method, after displaying the multimodal colposcopy image of the first time series, the medical worker determines the diagnosis result according to their own knowledge reserve, and the diagnosis result can display various selection items on the screen. For example, the selection items of the lesion size in the transformation zone include less than 5mm, 5-15mm or occupying 2 quadrants, and larger than 15mm or occupying 3-4 quadrants. The size is uncertain, and the user can select it according to the actual situation. The options for white epithelium include none or transparent, misty, translucent milky white, obvious, opaque white, and the user can select it according to the actual situation. The options for acetowhite epithelial surface and borders include: diffuse; clear but irregular borders, jagged, "geographical", satellite-like lesions; clear borders, uneven surface, including cuff-like gland openings, the user can tick according to the actual situation select. Finally, the proposed diagnosis result is determined according to all the contents checked by the user.

对于云端服务器智能识别的方式，在获取到按照第一时序采集到的阴道镜检查过程中的多模态阴道镜图像后，利用预先训练好的云端服务器的AI识别器进行拟诊结果识别，输出拟诊结果，能够快速、客观的得到拟诊结果。For the intelligent identification method of the cloud server, after acquiring the multimodal colposcopy images during the colposcopy process collected according to the first time sequence, the pre-trained AI identifier of the cloud server is used to identify the suspected diagnosis results, and output The proposed diagnosis results can be obtained quickly and objectively.

对于AI识别器的训练过程包括：从医疗结构获取多个检测者的多模态阴道镜图像以及对应的标识信息，该标识信息为专家的诊断结果；利用所有的多模态阴道镜图像以及对应的标识信息对模型进行训练，得到AI识别器。在实际应用的过程中也将检测得到的多模态阴道镜图像以及对应的标识进行存储，作为AI识别器再训练的训练样本，再次训练该识别器，提高其识别精度。进一步的，当AI识别器确定拟诊结果后，阴道镜人工智能辅助质控管理系统可以对与拟诊结果对应的区域进行可能性提示标识，以便提示医生重点观察该区域。The training process for the AI recognizer includes: obtaining the multimodal colposcopy images of multiple examiners and the corresponding identification information from the medical structure, and the identification information is the diagnosis result of the expert; using all the multimodal colposcopy images and corresponding identification information The identification information of the model is trained to obtain an AI recognizer. In the process of practical application, the detected multimodal colposcopy image and the corresponding logo are also stored as training samples for the retraining of the AI recognizer, and the recognizer is retrained to improve its recognition accuracy. Further, after the AI recognizer determines the proposed diagnosis result, the colposcopy artificial intelligence-assisted quality control management system can mark the area corresponding to the proposed diagnosis result with a possibility indication, so as to remind the doctor to focus on the area.

其中，S150，包括：确定评估总分所处的评估范围；根据转诊信息和评估范围，确定预测结果。Wherein, S150 includes: determining the evaluation range in which the total evaluation score is located; and determining the prediction result according to the referral information and the evaluation range.

请参考表2，表2为转诊信息、评估范围、预测结果的对应关系。Please refer to Table 2, which shows the correspondence between referral information, assessment scope, and predicted results.

根据评估总分以及相应的转诊信息确定预测结果，该对应关系是根据大量的样本数据得到的，作为参考提高了预测的准确性。The prediction result is determined according to the total evaluation score and the corresponding referral information. The corresponding relationship is obtained based on a large number of sample data, which improves the accuracy of the prediction as a reference.

表2转诊信息、评估范围、预测结果的对应关系表Table 2 Correspondence table of referral information, assessment scope, and predicted results

进一步的，阴道镜人工智能辅助质控管理方法，还包括：当患者对病变部位进行活检后，获取病变部位对应的活检结果；利用活检结果对预测结果进行复检，得到复检结果。Further, the colposcopy artificial intelligence-assisted quality control management method further includes: after the patient performs biopsy on the lesion, obtaining the biopsy result corresponding to the lesion; and rechecking the predicted result by using the biopsy result to obtain the recheck result.

在得到预测结果后，患者会对病变部位进行活检，并将活检结果上传至阴道镜人工智能辅助质控管理系统，在阴道镜人工智能辅助质控管理系统接收到该活检结果后，利用活检结果对预测结果进行自动复检，由于活检的准确性更高，因此将预测结果中与活检结果不同的地方进行修改，得到复检结果。该复检结果包括预测结果、修改后的预测结果。进一步的，会将修改后的预测结果与原始信息作为再次样本，当样本的数量达到设定数量时，对阴道镜人工智能辅助质控管理系统中预测结果的确定方式进行再次调整，以便提高预测的准确性。After obtaining the predicted result, the patient will biopsy the lesion and upload the biopsy result to the colposcopy AI-assisted quality control management system. After the colposcopy AI-assisted quality control management system receives the biopsy result, use the biopsy result Automatic re-examination is performed on the predicted results. Since the accuracy of the biopsy is higher, the differences between the predicted results and the biopsy results are modified to obtain the re-examination results. The recheck result includes the prediction result and the revised prediction result. Further, the revised prediction results and the original information will be used as re-samples. When the number of samples reaches the set number, the method of determining the prediction results in the colposcopy AI-assisted quality control management system will be re-adjusted in order to improve the prediction. accuracy.

进一步的，为了提高预测的准确性，本实施例还可以将相关数据传输至云端服务器，以便至少一位有经验的医疗工作者再次分析，得到阴道镜的分析结果，再反馈给云端服务器，实现多个医疗工作者同时会诊，提高诊断的准确性，相应的，阴道镜人工智能辅助质控管理方法，还包括：Further, in order to improve the accuracy of the prediction, in this embodiment, the relevant data can also be transmitted to the cloud server, so that at least one experienced medical worker can analyze it again, obtain the analysis result of the colposcopy, and then feed it back to the cloud server to realize Multiple medical workers consult at the same time to improve the accuracy of diagnosis. Correspondingly, the artificial intelligence-assisted quality control management method of colposcopy also includes:

将预测结果、所有多模态阴道镜图像以及转诊信息上传至云端服务器，以使云端服务器对应的医疗工作者进行人工分析，得到分析结果；Upload the prediction results, all multimodal colposcopy images and referral information to the cloud server, so that the medical workers corresponding to the cloud server can perform manual analysis to obtain the analysis results;

从云端服务器获取分析结果；Obtain analysis results from the cloud server;

根据分析结果对预测结果进行修正。The forecast results are revised based on the analysis results.

本实施例中在完成用户阴道镜的智能预测得到预测结果后，将预测结果和所有的多模态阴道镜图像以及转诊信息上传至云端服务器，实现多个医疗工作者同时会诊。当医疗工作者根据多模态阴道镜图像以及转诊信息同时参考预测结果进行分析，得到分析结果，该分析结果包括医疗工作者确定的病变类型和需要活检的病变部位。当得到医疗工作者的分析结果后，为了提高阴道镜检查的可靠性，会根据分析结果对预测结果进行修正，具体可以采用各项占比修正，或医疗工作者的职称高低修正，或医疗工作者的工作经验修正中的任意一种或者多种。In this embodiment, after completing the intelligent prediction of the user's colposcopy and obtaining the prediction result, the prediction result and all the multimodal colposcopy images and referral information are uploaded to the cloud server, so as to realize simultaneous consultation by multiple medical workers. When the medical worker analyzes the multimodal colposcopy image and the referral information while referring to the prediction result, the analysis result is obtained, and the analysis result includes the lesion type determined by the medical worker and the lesion site that needs to be biopsyed. When the analysis results of medical workers are obtained, in order to improve the reliability of colposcopy, the predicted results will be revised according to the analysis results. Any one or more of the modification of the work experience of the person.

由此，在一些实施例中，根据分析结果对预测结果进行修正，包括：根据多个分析结果中的各病变类型以及各需要活检的病变部位的占比进行确定；确定占比最高的病变类型和病变部位；利用占比最高的病变类型和病变部位对预测结果进行修正。Therefore, in some embodiments, revising the prediction result according to the analysis result includes: determining according to each lesion type in the multiple analysis results and the proportion of each lesion site requiring biopsy; determining the lesion type with the highest proportion and lesion location; the predictions were revised with the highest proportion of lesion type and lesion location.

在另一些实施例中，根据分析结果对预测结果进行修正，包括：确定职称最高的医疗工作者，当职称最高的医疗工作者的数量为一时，则根据职称最高的医疗工作者的分析结果对预测结果进行修正；当职称最高的医疗工作者的数量为多个时，则根据所有分析结果中的占比最高的病变类型和病变部位对预测结果进行修正，或，根据职称最高的医疗工作者中的工作经验最多的分析结果，对预测结果进行修正。In other embodiments, modifying the prediction result according to the analysis result includes: determining the medical worker with the highest professional title, and when the number of medical workers with the highest professional title is one, then adjusting the medical worker according to the analysis result of the medical worker with the highest professional title. The prediction results are revised; when the number of medical workers with the highest professional titles is more than one, the prediction results are revised according to the lesion type and lesion location with the highest proportion in all analysis results, or, according to the medical workers with the highest professional titles The results of the analysis with the most work experience in the program are revised, and the forecast results are revised.

在另一些实施例中，根据分析结果对预测结果进行修正，包括：确定分析结果对应的医疗工作者的工作经验；利用最高工作经验对应的分析结果，对预测结果进行修正。In other embodiments, revising the prediction result according to the analysis result includes: determining the work experience of the medical worker corresponding to the analysis result; and revising the prediction result by using the analysis result corresponding to the highest work experience.

基于上述任一实施例，为了能够简化阴道镜的操作，使经验不足的医疗工作者也可进行操作，阴道镜人工智能辅助质控管理方法，还包括：Based on any of the above-mentioned embodiments, in order to simplify the operation of the colposcope and enable inexperienced medical workers to perform the operation, the artificial intelligence-assisted quality control management method for colposcopy also includes:

实时检测是否具有与阴道镜人工智能辅助质控管理的当前阶段相匹配的提示信息，若具有，则在显示界面上显示提示信息。Real-time detection of whether there is prompt information that matches the current stage of colposcopy AI-assisted quality control management, and if so, the prompt information will be displayed on the display interface.

其中，可以将阴道镜人工智能辅助质控管理的阶段分为：多模态阴道镜图像采集、拟诊结果确定、转诊信息获取、预测；当然还可以包括：镜头参数质控、生理盐水阴道镜图像采集、可疑癌特征确定、评估信息显示、阴道镜检查条件充分性确定。Among them, the stages of colposcopy AI-assisted quality control management can be divided into: multimodal colposcopy image acquisition, determination of simulated diagnosis results, referral information acquisition, and prediction; of course, it can also include: lens parameter quality control, saline vaginal vaginal Endoscopy image acquisition, suspicious cancer feature determination, evaluation information display, and determination of colposcopy condition adequacy.

以多模态阴道镜图像采集为例，对应的提示信息可以包括：开始涂覆液体提醒、每张图像采集倒计时、提示保持镜头平稳。Taking multimodal colposcopy image acquisition as an example, the corresponding prompt information may include: a reminder to start applying liquid, a countdown for each image acquisition, and a prompt to keep the lens steady.

可见，通过对每阶段设置提示信息，以便人员能够根据提示信息执行对应的操作，或者了解当前所处的阶段，保证能够针对影响阴道镜产品质控管理的全环节因素和检查流程，更全面地规范操作流程要求，进行引导提示，保障采图质量，降低人为失误风险，提升基层诊疗水平。It can be seen that by setting prompt information for each stage, so that personnel can perform corresponding operations according to the prompt information, or understand the current stage, and ensure that the whole process of factors affecting the quality control management of colposcopy products and inspection procedures can be more comprehensively analyzed. Standardize the operation process requirements, provide guidance and prompts, ensure the quality of the map, reduce the risk of human error, and improve the level of grass-roots diagnosis and treatment.

本申请实施例提供的一种具体的阴道镜人工智能辅助质控管理的流程示意图，包括：A specific schematic flow chart of colposcopy artificial intelligence-assisted quality control management provided by the embodiments of the present application, including:

点击开始按钮进入阴道镜人工智能辅助控制管理程序；Click the start button to enter the colposcopy artificial intelligence-assisted control management program;

确定患者信息。判断是否为初次就诊，若是，则输入患者姓名、年龄，登记转诊阴道镜检查指征；若否，则从云端调入患者基本信息和转诊阴道镜检查指征。Determine patient information. It is judged whether it is the first visit. If so, enter the patient's name and age, and register the referral for colposcopy indication; if not, transfer the patient's basic information and referral colposcopy indication from the cloud.

进入观察检查阶段，暴露宫颈口，；Enter the observation and inspection stage, expose the cervix,;

调入阴道镜镜头的工作距离校准提示图片，此时AI智能分析功能未启动，最终能够得到阴道镜图像清晰的镜头参数。The working distance calibration prompt picture of the colposcopy lens is transferred. At this time, the AI intelligent analysis function is not activated, and finally the lens parameters with clear colposcopy images can be obtained.

距离校准完成后，调入原始图采集提示图片；After the distance calibration is completed, transfer the original image to collect the prompt image;

判断是否采集到第一张图；第一张图像当采用生理盐水清洁后，符合清晰度，放大倍数要求、目标物没有异物或者没有被异物遮挡；Determine whether the first image is collected; the first image, after being cleaned with normal saline, meets the requirements of clarity, magnification, and the target object is free of foreign objects or blocked by foreign objects;

检查流程中，获取多张阴道镜图像，对多张阴道镜图像进行动态识别，确定是否具有可疑癌基本六项特征；若是，则确定活检位置；可疑癌基本六项特征，包括：肿瘤或新生肿物、外生型病变、脆性血管、表面不规则、坏死、溃疡(坏死性)；无上述现象若否，则在显示界面显示待评估信息，待评估信息用于医护人员进行选择；待评估信息包括检查7项；需要都选择；宫颈上皮雌激素作用影响、绝经后上皮萎缩性改变，检查前是否做过雌激素治疗、妊娠新改变存在、阴道镜下确定急性感染存在(病毒、真菌、细菌)、镜下能充分观察到整个宫颈和阴道上段、确认是否接受过治疗或镜下可见治疗后改变、上皮纤维化或镜下不同程度可见。During the inspection process, multiple colposcopy images are obtained, and the multiple colposcopy images are dynamically identified to determine whether there are six basic characteristics of suspicious cancer; if so, determine the biopsy location; the six basic characteristics of suspicious cancer, including: tumor or new Tumors, exophytic lesions, fragile blood vessels, surface irregularities, necrosis, ulcers (necrosis); if none of the above phenomena are present, the information to be evaluated will be displayed on the display interface, and the to-be-evaluated information will be used by medical staff for selection; to be evaluated The information includes 7 examination items; need to choose all; the effect of cervical epithelial estrogen, postmenopausal epithelial atrophy, whether estrogen therapy has been done before the examination, the existence of new changes in pregnancy, the existence of acute infection (virus, fungus, Bacteria), the entire cervix and upper vaginal segment can be fully observed under the microscope, whether the treatment has been received or the changes after treatment can be seen under the microscope, epithelial fibrosis or different degrees of visible under the microscope.

接收到医护人员选择后的已评估信息；Received the assessed information as selected by the healthcare provider;

根据已评估信息确定阴道镜检查条件是否充分；Determine the adequacy of colposcopy based on assessed information;

若充分则当用醋酸涂覆整个宫颈后，按照计时器设定周期自动采集图像，同时，在显示界面上将采集的图像放大设定倍数，以确保可见完整的宫颈图像；同时当图不清楚时，用户可手动操作补图，不影响设定采集图像的程序；If it is sufficient, after coating the entire cervix with acetic acid, the image will be automatically collected according to the timer setting period, and at the same time, the collected image will be enlarged on the display interface by a set multiple to ensure that the complete cervical image can be seen; at the same time, when the image is unclear When , the user can manually operate the supplementary image, which does not affect the setting of the image acquisition program;

对按照时序采集的图像上传云服务进行分析，得到拟诊结果，并将拟诊结果反馈个阴道镜图像处理系统；Analyze the images collected according to the time sequence and upload the cloud service to obtain the proposed diagnosis result, and feed the proposed diagnosis result back to a colposcopy image processing system;

用户填写阴道镜检查图像所见(拟诊结果评分七项)；转化区病变大小、出现白上皮、醋白上皮表面和边界、血管、碘染色(白上皮对应处)、转化区类型、其它1，可见持续性出血，可见易脆组织，可见白斑涂醋酸前存在。The user fills in the findings of the colposcopy image (seven items of the proposed diagnosis result); the size of the lesion in the transformation zone, the appearance of white epithelium, the surface and border of the acetowhite epithelium, blood vessels, iodine staining (corresponding to the white epithelium), the type of the transformation zone, other 1 , visible persistent bleeding, visible brittle tissue, visible leukoplakia existed before acetic acid.

对于两个系统确定的拟诊和用户确定的拟诊，存在差异的进行在线标注和病例反馈，上传至云端存储至算法学习库，当达到设定量后，进行在学习更新。For the proposed diagnosis determined by the two systems and the proposed diagnosis determined by the user, online annotation and case feedback are carried out if there are differences, and they are uploaded to the cloud for storage to the algorithm learning library.

根据拟诊结果，利用预先设定好的评估表，进行评估，确定评估总分；According to the results of the proposed diagnosis, use a pre-set evaluation form to conduct an evaluation to determine the total evaluation score;

获取转诊信息，根据转诊信息以及评估总分，得到智能评估结果和智能评估提示。Obtain referral information, and obtain intelligent evaluation results and intelligent evaluation prompts based on the referral information and the total evaluation score.

上述实施例从方法流程的角度介绍一种阴道镜人工智能辅助质控管理方法，下述实施例从虚拟模块或者虚拟单元的角度介绍了一种阴道镜人工智能辅助质控管理装置，具体详见下述实施例。The above embodiment introduces a colposcopy artificial intelligence-assisted quality control management method from the perspective of the method process. The following embodiment introduces a colposcopy artificial intelligence-assisted quality control management device from the perspective of a virtual module or virtual unit. The following examples.

本申请实施例提供一种阴道镜人工智能辅助质控管理装置，如图5所示，该装置具体可以包括：The embodiment of the present application provides a colposcopy artificial intelligence-assisted quality control management device, as shown in FIG. 5 , the device may specifically include:

多模态阴道镜图像获取模块510，用于获取按照第一时序采集的阴道镜检查的过程中的多模态阴道镜图像，其中，所述阴道镜检查包括醋白检查，或，醋白检查和补碘检查；The multimodal colposcopyimage acquisition module 510 is configured to acquire the multimodal colposcopy images during the colposcopy process acquired according to the first sequence, wherein the colposcopy examination includes acetowhite examination, or acetowhite examination and iodine supplementation check;

拟诊模块520，用于根据所有的所述多模态阴道镜图像的变化情况，确定拟诊结果，其中，所述拟诊结果包括：转化区病变大小、出现白上皮、醋白上皮表面和边界、血管、转化区类型、其它信息的性状，所述其他信息的性状包括可见持续性出血，可见易脆组织，可见白斑涂醋酸前存在是否可见；当所述阴道镜检查包括醋白检查和补碘检查时，所述拟诊结果还包括碘染色的性状；Adiagnosis module 520 is configured to determine a diagnosis result according to the changes of all the multimodal colposcopy images, wherein the diagnosis result includes: the size of the lesion in the transformation zone, the appearance of white epithelium, the surface of acetowhite epithelium, and the Boundaries, blood vessels, types of transformation zones, and other informative traits including visible persistent bleeding, visible brittle tissue, visible vitiligo prior to acetic acid application or not; when the colposcopy examination includes acetowhite examination and During the iodine supplementation examination, the proposed diagnosis result also includes the characteristics of iodine staining;

评分模块530，用于按照预先设定的评估信息，对所述拟诊结果进行评分，得到评估总分；转诊信息获取模块540，用于获取转诊信息；其中，所述转诊信息包括：HPV检测信息、细胞学信息；Thescoring module 530 is used to score the proposed diagnosis result according to the preset evaluation information to obtain the total evaluation score; the referralinformation acquisition module 540 is used to acquire the referral information; wherein the referral information includes : HPV detection information, cytology information;

预测结果确定模块550，用于根据所述转诊信息和所述评估总分，确定预测结果，其中所述预测结果包括病变类型和病变部位活检提示。The predictionresult determination module 550 is configured to determine a prediction result according to the referral information and the total evaluation score, wherein the prediction result includes the lesion type and the biopsy prompt of the lesion site.

优选的，还包括：Preferably, it also includes:

活检结果获取模块，用于当患者对所述病变部位进行活检后，获取所述病变部位对应的活检结果；A biopsy result acquisition module, configured to obtain a biopsy result corresponding to the lesion site after the patient performs a biopsy on the lesion site;

复检模块，用于利用所述活检结果对所述预测结果进行复检，得到复检结果。A re-examination module, configured to re-examine the predicted result by using the biopsy result to obtain a re-examination result.

优选的，还包括：Preferably, it also includes:

上传模块，用于将所述预测结果、所有多模态阴道镜图像以及所述转诊信息上传至云端服务器，以使所述云端服务器对应的医疗工作者进行人工分析，得到分析结果；an uploading module for uploading the prediction results, all multimodal colposcopy images and the referral information to a cloud server, so that medical workers corresponding to the cloud server can perform manual analysis to obtain analysis results;

分析结果获取模块，用于从所述云端服务器获取所述分析结果；an analysis result obtaining module, configured to obtain the analysis result from the cloud server;

修正模块，用于根据所述分析结果对所述预测结果进行修正。A correction module, configured to correct the prediction result according to the analysis result.

优选的，拟诊模块520，包括：Preferably, thediagnosis module 520 includes:

第一拟诊单元，用于在显示界面按照第一时序显示所有的所述多模态阴道镜图像，以便当前医疗工作者根据所有的所述多模态阴道镜图像的变化情况输入拟诊结果；获取所述用户确定的所述拟诊结果；A first simulating unit, configured to display all the multimodal colposcopy images on the display interface according to a first time sequence, so that the current medical worker can input the simulating results according to the changes of all the multimodal colposcopy images ; Obtain the proposed diagnosis result determined by the user;

第一拟诊单元，用于按照第一时序发送所有的所述多模态阴道镜图像至云端服务器，以便所述云端服务器对所述多模态阴道镜图像进行自动特征识别，并根据特征的变化情况，生成所述拟诊结果；获取所述云端服务器发送的所述拟诊结果。The first diagnosing unit is configured to send all the multimodal colposcopy images to the cloud server according to the first sequence, so that the cloud server can automatically identify the features of the multimodal colposcopy images, and According to the changing situation, the proposed diagnosis result is generated; and the proposed diagnosis result sent by the cloud server is obtained.

优选的，还包括：Preferably, it also includes:

清晰度获取模块，用于获取镜头参数下的阴道镜图像的清晰度，所述镜头参数包括镜头位置，镜头的各光源的照度、均匀值以及色温；A sharpness acquisition module, used for acquiring the sharpness of the colposcopy image under lens parameters, the lens parameters including the lens position, the illuminance, uniform value and color temperature of each light source of the lens;

镜头参数发送模块，用于发送镜头参数调整指令至阴道镜镜头，以使所述阴道镜镜头调整镜头参数，使各光源下的所述阴道镜图像的清晰度至设定清晰度；a lens parameter sending module, configured to send a lens parameter adjustment instruction to the colposcope lens, so that the colposcope lens adjusts the lens parameters, so that the clarity of the colposcope image under each light source reaches the set clarity;

异物确定模块，用于确定设定清晰度的阴道镜图像是否存在异物；The foreign body determination module is used to determine whether there is foreign body in the colposcopy image of the set definition;

提示模块，用于若是，发送提示信息至显示设备，提示信息用于提示医护人员去除异物；控制采集模块，用于若否，则按照校准后的镜头参数控制阴道镜镜头进行多模态阴道图像采集。The prompt module is used to send prompt information to the display device if it is, and the prompt information is used to prompt the medical staff to remove the foreign body; the control acquisition module is used to control the colposcope lens to perform multi-modal vaginal images according to the calibrated lens parameters. collection.

优选的，还包括：Preferably, it also includes:

生理盐水阴道镜图像获取模块，用于当宫颈部位涂覆生理盐水后，获取按照第二时序采集的多张生理盐水涂覆后的生理盐水阴道镜图像；The normal saline colposcopy image acquisition module is used for acquiring a plurality of normal saline colposcopy images after the normal saline is coated on the cervix after the cervix is coated according to the second time sequence;

可疑癌特征确定模块，用于对所有所述生理盐水阴道镜图像进行特征识别，确定是否具有可疑癌特征；所述可疑癌特征包括：肿瘤或新生肿物、外生型病变、脆性血管、表面不规则、坏死、溃疡；The suspicious cancer feature determination module is used to perform feature recognition on all the saline colposcopy images to determine whether there are suspicious cancer features; the suspicious cancer features include: tumors or new tumors, exophytic lesions, fragile blood vessels, surface irregularity, necrosis, ulceration;

待评估信息显示模块，用于若不具有，则在显示界面显示待评估信息，待评估信息包括：宫颈上皮雌激素作用影响、绝经后上皮萎缩性改变，检查前是否做过雌激素治疗、妊娠新改变存在、阴道镜下确定急性感染存在、镜下能充分观察到整个宫颈和阴道上段、确认是否接受过治疗或镜下可见治疗后改变、上皮纤维化或镜下不同程度可见；To-be-evaluated information display module, used to display the to-be-evaluated information on the display interface if not available, the to-be-evaluated information includes: the effect of estrogen on the cervical epithelium, postmenopausal epithelial atrophy changes, whether estrogen therapy has been done before the examination, pregnancy The existence of new changes, the existence of acute infection confirmed by colposcopy, the entire cervix and upper vaginal segment can be fully observed under the microscope, whether it has been treated or the changes after treatment can be seen under the microscope, epithelial fibrosis or different degrees of visible under the microscope;

已评估信息接收模块，用于接收到当前医疗工作者从所述待评估信息中选择的已评估信息；确定模块，用于根据已评估信息确定阴道镜检查条件是否充分，若是，则执行所述获取按照第一时序采集的阴道镜检查的过程中的多模态阴道镜图像的步骤。The evaluated information receiving module is used to receive the evaluated information selected by the current medical worker from the information to be evaluated; the determining module is used to determine whether the colposcopy condition is sufficient according to the evaluated information, and if so, execute the described The step of acquiring a multimodal colposcopy image in the process of colposcopy acquired according to the first time sequence.

优选的，还包括：Preferably, it also includes:

提示模块，用于实时检测是否具有与所述阴道镜人工智能辅助质控管理的当前阶段相匹配的提示信息，若具有，则在显示界面上显示所述提示信息。The prompt module is used to detect in real time whether there is prompt information matching the current stage of the colposcopy artificial intelligence-assisted quality control management, and if so, display the prompt information on the display interface.

优选的，预测结果确定模块550，包括：Preferably, the predictionresult determination module 550 includes:

评估范围确定单元，用于确定所述评估总分所处的评估范围；an evaluation range determination unit, used to determine the evaluation range in which the total evaluation score is located;

预测结果确定单元，用于根据所述转诊信息和所述评估范围，确定预测结果。A prediction result determination unit, configured to determine a prediction result according to the referral information and the evaluation range.

优选的，多模态阴道镜图像获取模块510，包括：Preferably, the multimodal colposcopyimage acquisition module 510 includes:

获取单元，用于获取按照第一时序采集的阴道镜检查的过程中的多模态阴道镜图像；an acquisition unit, configured to acquire the multimodal colposcopy images during the colposcopy examination collected according to the first sequence;

漏检判断单元，用于判断当前时刻是否存在多模态阴道镜图像的漏检情况；A missed detection judging unit, used for judging whether there is a missed detection situation of the multimodal colposcopy image at the current moment;

漏检提示与再采集单元，用于若存在当确定当前所述多模态阴道镜图像为模糊图像，则进行漏检提示，并控制图像采集设备进行再次采集，得到补充多模态阴道镜图像；The missed detection prompt and re-acquisition unit is used to prompt the missed detection when it is determined that the current multimodal colposcopy image is a blurred image, and control the image acquisition device to perform re-acquisition to obtain a supplementary multimodal colposcopy image ;

替换单元，用于根据所述标记将所述模糊图像替换成所述补充多模态阴道镜图像。A replacement unit for replacing the blurred image with the supplemental multimodal colposcopy image according to the marker.

本申请实施例中提供了一种阴道镜人工智能辅助质控管理系统，如图6所示，图6所示的阴道镜人工智能辅助质控管理系统600包括：处理器601和存储器603。其中，处理器601和存储器603相连，如通过总线602相连。可选地，阴道镜人工智能辅助质控管理系统600还可以包括收发器604。需要说明的是，实际应用中收发器604不限于一个，该阴道镜人工智能辅助质控管理系统600的结构并不构成对本申请实施例的限定。The embodiments of the present application provide a colposcopy artificial intelligence-assisted quality control management system, as shown in FIG. 6 , the colposcopy artificial intelligence-assisted qualitycontrol management system 600 shown in FIG. Theprocessor 601 is connected to thememory 603 , for example, through abus 602 . Optionally, the colposcopy AI-assisted qualitycontrol management system 600 may further include atransceiver 604 . It should be noted that, in practical applications, thetransceiver 604 is not limited to one, and the structure of the colposcopy artificial intelligence-assisted qualitycontrol management system 600 does not constitute a limitation to the embodiments of the present application.

处理器601可以是CPU(Central Processing Unit，中央处理器)，通用处理器，DSP(Digital Signal Processor，数据信号处理器)，ASIC(Application SpecificIntegrated Circuit，专用集成电路)，FPGA(Field Programmable Gate Array，现场可编程门阵列)或者其他可编程逻辑器件、晶体管逻辑器件、硬件部件或者其任意组合。其可以实现或执行结合本申请公开内容所描述的各种示例性的逻辑方框，模块和电路。处理器601也可以是实现计算功能的组合，例如包含一个或多个微处理器组合，DSP和微处理器的组合等。Theprocessor 601 may be a CPU (Central Processing Unit, central processing unit), a general-purpose processor, a DSP (Digital Signal Processor, data signal processor), an ASIC (Application Specific Integrated Circuit, an application-specific integrated circuit), an FPGA (Field Programmable Gate Array, Field Programmable Gate Array) or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It may implement or execute the various exemplary logical blocks, modules and circuits described in connection with this disclosure. Theprocessor 601 may also be a combination that realizes computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, and the like.

总线602可包括一通路，在上述组件之间传送信息。总线602可以是PCI(Peripheral Component Interconnect，外设部件互连标准)总线或EISA(ExtendedIndustry Standard Architecture，扩展工业标准结构)总线等。总线602可以分为地址总线、数据总线、控制总线等。为便于表示，图6中仅用一条粗线表示，但并不表示仅有一根总线或一种类型的总线。Thebus 602 may include a path to transfer information between the components described above. Thebus 602 may be a PCI (Peripheral Component Interconnect, Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture, Extended Industry Standard Architecture) bus or the like. Thebus 602 can be divided into an address bus, a data bus, a control bus, and the like. For ease of presentation, only one thick line is used in FIG. 6, but it does not mean that there is only one bus or one type of bus.

存储器603可以是ROM(Read Only Memory，只读存储器)或可存储静态信息和指令的其他类型的静态存储设备，RAM(Random Access Memory，随机存取存储器)或者可存储信息和指令的其他类型的动态存储设备，也可以是EEPROM(Electrically ErasableProgrammable Read Only Memory，电可擦可编程只读存储器)、CD-ROM(Compact DiscRead Only Memory，只读光盘)或其他光盘存储、光碟存储(包括压缩光碟、激光碟、光碟、数字通用光碟、蓝光光碟等)、磁盘存储介质或者其他磁存储设备、或者能够用于携带或存储具有指令或数据结构形式的期望的程序代码并能够由计算机存取的任何其他介质，但不限于此。Thememory 603 may be a ROM (Read Only Memory, read only memory) or other types of static storage devices that can store static information and instructions, a RAM (Random Access Memory, random access memory) or other types that can store information and instructions. A dynamic storage device can also be an EEPROM (Electrically Erasable Programmable Read Only Memory), a CD-ROM (Compact Disc Read Only Memory, a CD-ROM), or other CD-ROM storage, CD-ROM storage (including compressed CDs, Laser Disc, Optical Disc, Digital Versatile Disc, Blu-ray Disc, etc.), magnetic disk storage medium or other magnetic storage device, or any other capable of carrying or storing desired program code in the form of instructions or data structures and capable of being accessed by a computer medium, but not limited to this.

存储器603用于存储执行本申请方案的应用程序代码，并由处理器601来控制执行。处理器601用于执行存储器603中存储的应用程序代码，以实现前述方法实施例所示的内容。Thememory 603 is used to store the application code for executing the solution of the present application, and the execution is controlled by theprocessor 601 . Theprocessor 601 is configured to execute the application program code stored in thememory 603, so as to realize the content shown in the foregoing method embodiments.

其中，阴道镜人工智能辅助质控管理系统包括但不限于：移动电话、笔记本电脑、数字广播接收器、PDA(个人数字助理)、PAD(平板电脑)、PMP(便携式多媒体播放器)、车载终端(例如车载导航终端)等的移动终端以及诸如数字TV、台式计算机等的固定终端。还可以为服务器等。图6示出的阴道镜人工智能辅助质控管理系统仅是一个示例，不应对本公开实施例的功能和使用范围带来任何限制。Among them, the colposcopy artificial intelligence-assisted quality control management system includes but is not limited to: mobile phone, notebook computer, digital broadcast receiver, PDA (personal digital assistant), PAD (tablet computer), PMP (portable multimedia player), vehicle terminal Mobile terminals such as in-vehicle navigation terminals, etc., and stationary terminals such as digital TVs, desktop computers, and the like. It can also be a server, etc. The colposcopy artificial intelligence-assisted quality control management system shown in FIG. 6 is only an example, and should not impose any limitations on the functions and scope of use of the embodiments of the present disclosure.

本申请实施例提供了一种计算机可读存储介质，该计算机可读存储介质上存储有计算机程序，当其在计算机上运行时，使得计算机可以执行前述方法实施例中相应内容。Embodiments of the present application provide a computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and when the computer program runs on the computer, the computer can execute the corresponding content in the foregoing method embodiments.

应该理解的是，虽然附图的流程图中的各个步骤按照箭头的指示依次显示，但是这些步骤并不是必然按照箭头指示的顺序依次执行。除非本文中有明确的说明，这些步骤的执行并没有严格的顺序限制，其可以以其他的顺序执行。而且，附图的流程图中的至少一部分步骤可以包括多个子步骤或者多个阶段，这些子步骤或者阶段并不必然是在同一时刻执行完成，而是可以在不同的时刻执行，其执行顺序也不必然是依次进行，而是可以与其他步骤或者其他步骤的子步骤或者阶段的至少一部分轮流或者交替地执行。It should be understood that although the various steps in the flowchart of the accompanying drawings are sequentially shown in the order indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless explicitly stated herein, the execution of these steps is not strictly limited to the order and may be performed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but may be executed at different times, and the execution sequence is also It does not have to be performed sequentially, but may be performed alternately or alternately with other steps or at least a portion of sub-steps or stages of other steps.

以上所述仅是本申请的部分实施方式，应当指出，对于本技术领域的普通技术人员来说，在不脱离本申请原理的前提下，还可以做出若干改进和润饰，这些改进和润饰也应视为本申请的保护范围。The above are only part of the embodiments of the present application. It should be pointed out that for those skilled in the art, without departing from the principles of the present application, several improvements and modifications can also be made. It should be regarded as the protection scope of this application.

Claims (10)

1. A colposcope artificial intelligence auxiliary quality control management method is characterized by comprising the following steps:

acquiring a multi-modal colposcopic image in the process of colposcopy acquired according to a first time sequence, wherein the colposcopy comprises a vinegar white examination, or a vinegar white examination and an iodine supplementation examination;

determining a diagnosis simulating result according to the change condition of all the multi-modal colposcopic images, wherein the diagnosis simulating result comprises: transformation zone lesion size, appearance of white epithelium, vinegar-white epithelium surface and boundary, blood vessel, transformation zone type, and properties of other information including visible persistent hemorrhage, visible fragile tissue, and visible white spot presence or absence before coating acetic acid; when the colposcopy comprises a vinegar white examination and an iodine supplement examination, the diagnosis simulating result further comprises the character of iodine staining;

scoring the result to be diagnosed according to preset assessment information to obtain an assessment total score;

acquiring referral information; wherein the referral information comprises: HPV detection information, cytology information;

and determining a prediction result according to the referral information and the assessment total score, wherein the prediction result comprises a lesion type and a lesion site biopsy prompt.

2. The colposcope artificial intelligence auxiliary quality control management method according to claim 1, further comprising:

when a patient performs biopsy on the lesion part, acquiring a biopsy result corresponding to the lesion part;

and rechecking the prediction result by using the biopsy result to obtain a rechecking result.

3. The colposcope artificial intelligence auxiliary quality control management method according to claim 1, further comprising:

uploading the prediction result, all multi-modal colposcope images and referral information to a cloud server so that medical workers corresponding to the cloud server perform manual analysis to obtain an analysis result;

obtaining the analysis result from the cloud server;

and correcting the prediction result according to the analysis result.

4. The colposcope artificial intelligence auxiliary quality control management method according to claim 1, wherein the determining a diagnosis planning result according to the change condition of all the multi-modal colposcopic images comprises:

displaying all the multi-modal colposcopic images on a display interface according to a first time sequence so that a current medical worker inputs a diagnosis simulation result according to the change condition of all the multi-modal colposcopic images;

acquiring the result of the diagnosis planning determined by the user;

or the like, or, alternatively,

sending all the multi-modal colposcope images to a cloud server according to a first time sequence so that the cloud server can automatically identify the characteristics of the multi-modal colposcope images and generate the diagnosis simulation result according to the change condition of the characteristics;

and acquiring the diagnosis planning result sent by the cloud server.

5. The colposcopic artificial intelligence-aided quality control management method according to claim 1, wherein before the obtaining of the multi-modal colposcopic image in the process of the colposcopic examination collected according to the first time sequence, the method further comprises:

acquiring the definition of a colposcope image under lens parameters, wherein the lens parameters comprise a lens position, and the illumination, the uniform value and the color temperature of each light source of a lens;

sending a lens parameter adjusting instruction to a colposcope lens so that the colposcope lens can adjust lens parameters and the definition of the colposcope image under each light source can reach a set definition;

determining whether foreign matters exist in the colposcope image with set definition;

if so, sending prompt information to the display device, wherein the prompt information is used for prompting medical personnel to remove the foreign matter;

if not, controlling the colposcope lens according to the calibrated lens parameters to acquire the multi-mode vaginal image.

6. The colposcope artificial intelligence auxiliary quality control management method according to claim 5, further comprising:

after the cervix uteri part is coated with the normal saline, acquiring a plurality of normal saline colposcopic images which are acquired according to a second time sequence and coated with the normal saline;

performing feature recognition on all the physiological saline colposcopic images to determine whether suspicious cancer features exist; the suspected cancer features include: tumors or new tumors, extrinsic lesions, fragile vessels, surface irregularities, necrosis, ulcers;

if not, displaying the information to be evaluated on a display interface, wherein the information to be evaluated comprises: effects of estrogen on cervical epithelium, atrophic changes in epithelium after menopause, whether estrogen treatment was done before examination, the presence of new changes in pregnancy, the presence of acute infection under colposcopy, full observation of the entire cervical and vaginal upper segments under the scope, confirmation of whether changes were done after treatment or seen under the scope, epithelial fibrosis, or varying degrees under the scope;

receiving evaluated information selected by a current medical worker from the information to be evaluated;

and determining whether the colposcopy examination condition is sufficient according to the evaluated information, and if so, executing the step of acquiring the multi-mode colposcopy image acquired according to the first time sequence in the colposcopy examination process.

7. The colposcope artificial intelligence auxiliary quality control management method according to claim 1, further comprising:

and detecting whether prompt information matched with the current stage of the colposcope artificial intelligence auxiliary quality control management exists or not in real time, and if so, displaying the prompt information on a display interface.

8. The colposcope artificial intelligence auxiliary quality control management method according to claim 1, wherein the determining a prediction result according to the referral information and the assessment total score, wherein the prediction result comprises a lesion type and a lesion biopsy prompt, comprises:

determining an evaluation range in which the evaluation total score is positioned;

and determining a prediction result according to the referral information and the evaluation range.

9. The colposcopic artificial intelligence-aided quality control management method as claimed in claim 1, wherein the acquiring of the multi-modal colposcopic image in the process of the colposcopic examination collected according to the first time sequence comprises:

acquiring a multi-modal colposcopic image acquired according to a first time sequence in the colposcopic examination process;

judging whether the multi-mode colposcope image is missed at the current moment;

if the current multi-modal colposcope image is determined to be a fuzzy image, performing omission detection prompting, and controlling an image acquisition device to perform secondary acquisition to obtain a supplementary multi-modal colposcope image;

replacing the blurred image with the supplemental multi-modal colposcopic image according to the indicia.

10. The utility model provides a supplementary quality control management device of colposcope artificial intelligence which characterized in that includes:

the system comprises a multi-modal colposcope image acquisition module, a data acquisition module and a data processing module, wherein the multi-modal colposcope image acquisition module is used for acquiring a multi-modal colposcope image in a colposcopy process acquired according to a first time sequence, and the colposcopy comprises a vinegar white examination or a vinegar white examination and an iodine supplement examination;

the virtual diagnosis module is used for determining a virtual diagnosis result according to the change conditions of all the multi-modal colposcopic images, wherein the virtual diagnosis result comprises: transformation zone lesion size, appearance of white epithelium, vinegar-white epithelium surface and boundary, blood vessel, transformation zone type, and properties of other information including visible persistent hemorrhage, visible fragile tissue, and visible white spot presence or absence before coating acetic acid; when the colposcopy comprises a vinegar white examination and an iodine supplement examination, the diagnosis simulating result further comprises the character of iodine staining;

the scoring module is used for scoring the result to be diagnosed according to preset assessment information to obtain an assessment total score;

the referral information acquisition module is used for acquiring referral information; wherein the referral information comprises: HPV detection information, cytology information;

and the prediction result determining module is used for determining a prediction result according to the referral information and the assessment total score, wherein the prediction result comprises a lesion type and a lesion site biopsy prompt.

Images