技术领域Technical field
本公开涉及医学诊断领域，具体涉及一种医学诊断装置、电子设备及存储介质。The present disclosure relates to the field of medical diagnosis, and specifically to a medical diagnosis device, electronic equipment and storage medium.
背景技术Background technique
通常，医学影像识别技术或医疗大数据分析可利用人工智能在检测、分类和深度学习的技术优势，帮助医生更快获取信息，进行定量分析，成为协助医生完成诊断、治疗工作的一种辅助工具。Usually, medical image recognition technology or medical big data analysis can take advantage of the technical advantages of artificial intelligence in detection, classification and deep learning to help doctors obtain information faster and conduct quantitative analysis, becoming an auxiliary tool to assist doctors in completing diagnosis and treatment. .
卷积神经网络等机器学习算法凭借其自动特征学习、高精度、可扩展性强等优势在医学影像识别、基于医疗知识的智能辅助诊断等领域脱颖而出。相关技术中的一种医学影像识别方法，通过提取目标组织的核磁共振影像的一阶统计量、灰度共生矩阵、灰度游程矩阵、灰度区域大小矩阵、邻域灰度差矩阵等特征，并利用BP神经网络等机器学习算法，开展目标组织的病变等级诊断；本方法通过筛选有限数量的特征，降低了算法的计算难度，但其特征和算法的可解释性差，输出结果难以可视化，导致医生面对输出结果时，难以从中得知其判断依据，进而妨碍医生针对输出结果进行验证和定量分析。另一相关技术中的一种肺部病变良恶性风险分层辅助诊断系统，包括结节检测、医生指导、语义标注生成、样本生成、在线学习等步骤；在诊断过程中，需要医生标注关注点，计算得到关注区域内的肺结节轮廓与轮廓内各像素点的预测概率，因此，本方法虽然能够识别出肺结节的具体轮廓和概率，但识别过程需要人工干预，诊断效率较低。同时，上述方法均针对同一模态数据进行识别诊断，对跨模态数据，如文本、数值等不同模态数据的应用存在困难。Machine learning algorithms such as convolutional neural networks stand out in fields such as medical image recognition and intelligent auxiliary diagnosis based on medical knowledge due to their advantages such as automatic feature learning, high accuracy, and strong scalability. A medical image recognition method in the related art, by extracting first-order statistics, gray level co-occurrence matrix, gray level run length matrix, gray level area size matrix, neighborhood gray level difference matrix and other features of the MRI image of the target tissue, And use machine learning algorithms such as BP neural network to carry out lesion level diagnosis of target tissues; this method reduces the computational difficulty of the algorithm by screening a limited number of features, but its features and algorithms have poor interpretability, and the output results are difficult to visualize, resulting in When doctors face the output results, it is difficult to know the basis for their judgment, which hinders doctors from verifying and quantitatively analyzing the output results. Another related technology is an auxiliary diagnosis system for benign and malignant risk stratification of lung lesions, including nodule detection, doctor guidance, semantic annotation generation, sample generation, online learning and other steps; during the diagnosis process, doctors need to mark points of concern , calculate the contour of the pulmonary nodule in the area of interest and the predicted probability of each pixel within the contour. Therefore, although this method can identify the specific contour and probability of the pulmonary nodule, the identification process requires manual intervention and the diagnostic efficiency is low. At the same time, the above methods all perform identification and diagnosis on the same modal data, and it is difficult to apply cross-modal data, such as text, numerical data and other different modal data.
发明内容Contents of the invention
有鉴于此，本公开提供一种医学诊断装置、电子设备及存储介质，以解决跨模态数据协同应用以及相关技术中输出人工智能输出结果可解释性不好的问题。In view of this, the present disclosure provides a medical diagnosis device, electronic equipment and storage medium to solve the problem of poor interpretability of artificial intelligence output results in cross-modal data collaborative applications and related technologies.
第一方面，提供一种医学诊断装置，包括：第一获取模块，用于获取医学影像中的目标结构的第一异常特征信息，第一异常特征信息表征目标结构是否出现异常；第二获取模块，用于获取目标结构对应的疾病类型列表，疾病类型列表中包括发生于目标结构的多个疾病；诊断模块，用于基于第一异常特征信息和多个疾病之间的关联度，得到第一诊断信息。In a first aspect, a medical diagnosis device is provided, including: a first acquisition module for acquiring first abnormal feature information of a target structure in a medical image, where the first abnormal feature information represents whether the target structure is abnormal; a second acquisition module , used to obtain a list of disease types corresponding to the target structure. The disease type list includes multiple diseases that occur in the target structure; the diagnosis module is used to obtain the first abnormal feature information and the correlation between the multiple diseases. diagnostic information.
在一些实施例中，基于第一异常特征信息和多个疾病之间的关联度，得到第一诊断信息，包括：遍历疾病类型列表中的多个疾病；根据目标结构对应的专家意见信息，计算第一异常特征信息与每个疾病之间的第一关联度；根据目标结构对应的历史诊断信息，计算第一异常特征信息与每个疾病之间的第二关联度；基于第一异常特征信息与每个疾病之间的第一关联度和第二关联度，得到第一诊断信息。In some embodiments, obtaining the first diagnostic information based on the correlation between the first abnormal feature information and multiple diseases includes: traversing multiple diseases in the disease type list; calculating based on expert opinion information corresponding to the target structure. The first degree of correlation between the first abnormal characteristic information and each disease; calculating the second degree of correlation between the first abnormal characteristic information and each disease based on the historical diagnosis information corresponding to the target structure; based on the first abnormal characteristic information The first degree of correlation and the second degree of correlation with each disease are used to obtain first diagnostic information.
在一些实施例中，基于第一异常特征信息与每个疾病之间的第一关联度和第二关联度，得到第一诊断信息，包括：基于第一异常特征信息与每个疾病之间的第一关联度，得到第二诊断信息，第二诊断信息中包括至少一个疾病，以及诊断为至少一个疾病中每个疾病的概率；基于第一异常特征信息与每个疾病之间的第二关联度，得到第三诊断信息，第三诊断信息包括至少一个疾病，以及诊断为至少一个疾病中每个疾病的概率；基于第二诊断信息和第三诊断信息中每个疾病的概率，得到第一诊断信息。In some embodiments, obtaining the first diagnosis information based on the first correlation degree and the second correlation degree between the first abnormal characteristic information and each disease includes: based on the first correlation degree between the first abnormal characteristic information and each disease. First degree of correlation, obtain second diagnostic information, the second diagnostic information includes at least one disease, and the probability of diagnosing each of the at least one disease; based on the second correlation between the first abnormal feature information and each disease degree, obtain third diagnostic information, the third diagnostic information includes at least one disease, and the probability of diagnosis of each disease in at least one disease; based on the second diagnostic information and the probability of each disease in the third diagnostic information, obtain the first diagnostic information.
在一些实施例中，基于第一异常特征信息与每个疾病之间的第一关联度，得到第二诊断信息，包括：将疾病类型列表中的多个疾病按照第一关联度由大到小的顺序排列，得到疾病队列；基于疾病队列，生成第二诊断信息。In some embodiments, obtaining the second diagnostic information based on the first degree of correlation between the first abnormal characteristic information and each disease includes: sorting multiple diseases in the disease type list from large to small according to the first degree of correlation. Arrange in order to obtain a disease queue; based on the disease queue, generate second diagnostic information.
在一些实施例中，专家意见信息包括原始异常样本数据和与原始异常样本数据对应的专家诊断结果，专家诊断结果由德尔菲法得到；其中，根据目标结构对应的专家意见信息，计算第一异常特征信息与每个疾病之间的第一关联度，包括：基于原始异常样本数据，得到第二异常特征信息；基于第二异常特征信息和专家诊断结果，得到疾病类型列表中每个疾病与患者体征之间的相关性；利用注意力机制，基于每个疾病与患者体征之间的相关性和第一异常特征信息，得到第一异常特征信息与每个疾病之间的第一关联度。In some embodiments, the expert opinion information includes original abnormal sample data and expert diagnosis results corresponding to the original abnormal sample data. The expert diagnosis results are obtained by the Delphi method; wherein, according to the expert opinion information corresponding to the target structure, the first abnormality is calculated The first degree of correlation between feature information and each disease includes: based on the original abnormal sample data, obtaining the second abnormal feature information; based on the second abnormal feature information and expert diagnosis results, obtaining each disease and patient in the disease type list Correlation between signs; using the attention mechanism, based on the correlation between each disease and the patient's signs and the first abnormal feature information, the first correlation degree between the first abnormal feature information and each disease is obtained.
在一些实施例中，历史诊断信息包括第三异常特征信息和与第三异常特征信息对应的历史诊断结果，根据目标结构对应的历史诊断信息，计算第一异常特征信息与每个疾病之间的第二关联度，包括：计算第三异常特征信息和历史诊断结果间的频繁项集；基于频繁项集，得到疾病类型列表中每个疾病与患者体征之间的相关性；利用注意力机制，基于每个疾病与患者体征之间的相关性和第一异常特征信息，得到第一异常特征信息与每个疾病之间的第二关联度。In some embodiments, the historical diagnosis information includes third abnormal feature information and historical diagnosis results corresponding to the third abnormal feature information. According to the historical diagnosis information corresponding to the target structure, the relationship between the first abnormal feature information and each disease is calculated. The second degree of correlation includes: calculating the frequent item sets between the third abnormal feature information and historical diagnosis results; based on the frequent item sets, obtaining the correlation between each disease in the disease type list and the patient's signs; using the attention mechanism, Based on the correlation between each disease and the patient's physical signs and the first abnormal feature information, a second degree of correlation between the first abnormal feature information and each disease is obtained.
在一些实施例中，诊断模块还用于：将第一异常特征信息输入预测模型，得到第四诊断信息；基于第一诊断信息和第四诊断信息，得到第五诊断信息；其中，预测模型包括随机森林模型。In some embodiments, the diagnosis module is also configured to: input the first abnormal characteristic information into the prediction model to obtain fourth diagnostic information; and obtain fifth diagnostic information based on the first diagnostic information and the fourth diagnostic information; wherein the prediction model includes Random forest model.
在一些实施例中，获取医学影像中的目标结构的第一异常特征信息，包括：基于医学影像提取影像异常特征信息，影像异常特征信息包括：目标结构的形态信息、目标结构图像的灰度信息、目标结构的尺寸信息和目标结构的空间位置信息；和/或获取与医学影像对应的病历文本，并基于病历文本提取文本异常特征信息；和/或获取与医学影像对应的数值性检查报告，并基于数值性检查报告提取数值异常特征信息；基于影像异常特征信息、文本异常特征信息、数值异常特征信息，确定第一异常特征信息。In some embodiments, obtaining the first abnormal feature information of the target structure in the medical image includes: extracting image abnormal feature information based on the medical image. The image abnormal feature information includes: morphological information of the target structure, and grayscale information of the target structure image. , the size information of the target structure and the spatial position information of the target structure; and/or obtain the medical record text corresponding to the medical image, and extract text abnormal feature information based on the medical record text; and/or obtain the numerical inspection report corresponding to the medical image, And extract the numerical abnormality feature information based on the numerical inspection report; determine the first abnormality feature information based on the image abnormality feature information, text abnormality feature information, and numerical abnormality feature information.
在一些实施例中，医学诊断装置应用于中耳病变诊断，第一异常特征信息由中耳图像提取，包括：中耳形态信息、中耳图像的灰度信息、中耳尺寸信息和中耳空间位置信息。In some embodiments, the medical diagnostic device is applied to the diagnosis of middle ear lesions, and the first abnormal feature information is extracted from the middle ear image, including: middle ear morphological information, grayscale information of the middle ear image, middle ear size information, and middle ear space. location information.
在一些实施例中，医学诊断装置应用于肝脏疾病诊断，第一异常特征信息由肝脏图像提取，包括：肝脏形态信息、肝脏图像的灰度信息、肝脏尺寸信息和肝脏空间位置信息。In some embodiments, the medical diagnostic device is applied to liver disease diagnosis, and the first abnormal feature information is extracted from the liver image, including: liver morphological information, grayscale information of the liver image, liver size information, and liver spatial location information.
第二方面，提供一种电子设备，包括：处理器；以及存储器，用于存储处理器的可执行指令；其中，处理器配置为经由执行可执行指令来执行一种医学诊断方法，医学诊断方法，包括：获取医学影像中的目标结构的第一异常特征信息，第一异常特征信息表征目标结构是否出现异常；获取目标结构对应的疾病类型列表，疾病类型列表中包括发生于目标结构的多个疾病；基于第一异常特征信息和多个疾病之间的关联度，得到第一诊断信息。In a second aspect, an electronic device is provided, including: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to execute a medical diagnosis method by executing the executable instructions, and the medical diagnosis method , including: obtaining the first abnormal feature information of the target structure in the medical image, and the first abnormal feature information represents whether the target structure is abnormal; obtaining a disease type list corresponding to the target structure, and the disease type list includes multiple diseases that occur in the target structure Disease; based on the correlation between the first abnormal characteristic information and multiple diseases, the first diagnostic information is obtained.
第三方面，提供一种计算机可读存储介质，其上存储有计算机程序，计算机程序被处理器执行时实现一种医学诊断方法，医学诊断方法，包括：获取医学影像中的目标结构的第一异常特征信息，第一异常特征信息表征目标结构是否出现异常；获取目标结构对应的疾病类型列表，疾病类型列表中包括发生于目标结构的多个疾病；基于第一异常特征信息和多个疾病之间的关联度，得到第一诊断信息。In a third aspect, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, a medical diagnosis method is implemented. The medical diagnosis method includes: obtaining a first step of a target structure in a medical image. Abnormal feature information, the first abnormal feature information represents whether the target structure is abnormal; obtain a disease type list corresponding to the target structure, the disease type list includes multiple diseases that occur in the target structure; based on the first abnormal feature information and the multiple diseases degree of correlation between them to obtain the first diagnostic information.
本公开实施例提供的医学诊断装置，能够获取医学影像中的目标结构的第一异常特征信息以及目标结构对应的疾病类型列表，根据第一异常特征信息和多个疾病之间的关联度，得到第一诊断信息。其中，第一诊断信息是基于第一异常特征信息中各结构特征得到的，因此，根据本实施例装置的输出结果能够清晰地确定异常特征信息和疾病之间的对应关系，有着良好的可解释性；同时，用户在面对输出结果时，能够根据目标结构的结构特征，了解疾病的判断依据，以验证输出结果的可信度。The medical diagnosis device provided by the embodiment of the present disclosure can obtain the first abnormal feature information of the target structure in the medical image and the list of disease types corresponding to the target structure. According to the correlation between the first abnormal feature information and multiple diseases, we obtain First diagnostic information. Among them, the first diagnostic information is obtained based on each structural feature in the first abnormal feature information. Therefore, the output result of the device according to this embodiment can clearly determine the corresponding relationship between the abnormal feature information and the disease, and has good interpretability. At the same time, when faced with the output results, users can understand the basis for disease judgment based on the structural characteristics of the target structure to verify the credibility of the output results.
附图说明Description of drawings
图1所示为本公开实施例提供的一种医学诊断装置所适用的场景示意图。FIG. 1 shows a schematic diagram of a scene applicable to a medical diagnosis device provided by an embodiment of the present disclosure.
图2所示为本公开实施例提供的一种医学诊断装置的结构示意图。FIG. 2 shows a schematic structural diagram of a medical diagnostic device provided by an embodiment of the present disclosure.
图3所示为本公开实施例提供的一种基于第一异常特征信息和多个疾病之间的关联度，得到第一诊断信息的流程示意图。FIG. 3 shows a schematic flowchart of obtaining first diagnostic information based on the correlation between first abnormal feature information and multiple diseases provided by an embodiment of the present disclosure.
图4所示为本公开实施例提供的一种根据目标结构对应的专家意见信息，计算第一异常特征信息与每个疾病之间的第一关联度的流程示意图。Figure 4 shows a schematic flowchart of calculating the first degree of correlation between the first abnormal feature information and each disease based on expert opinion information corresponding to the target structure provided by an embodiment of the present disclosure.
图5所示为本公开实施例提供的一种根据目标结构对应的历史诊断信息，计算第一异常特征信息与每个疾病之间的第二关联度的流程示意图。FIG. 5 shows a schematic flowchart of calculating the second degree of correlation between the first abnormal feature information and each disease based on the historical diagnosis information corresponding to the target structure provided by an embodiment of the present disclosure.
图6所示为本公开实施例提供的一种基于第一异常特征信息与每个疾病之间的第一关联度和第二关联度，得到第一诊断信息的流程示意图。Figure 6 shows a schematic flowchart of obtaining first diagnostic information based on the first degree of correlation and the second degree of correlation between the first abnormal feature information and each disease provided by an embodiment of the present disclosure.
图7所示为本公开实施例提供的一种诊断模块所实施步骤的流程示意图。FIG. 7 is a schematic flowchart of steps implemented by a diagnostic module provided by an embodiment of the present disclosure.
图8所示为本公开实施例提供的一种获取医学影像中的目标结构的第一异常特征信息的流程示意图。FIG. 8 shows a schematic flowchart of obtaining first abnormal feature information of a target structure in a medical image according to an embodiment of the present disclosure.
图9所示为本公开实施例提供的一种电子设备的结构示意图。FIG. 9 shows a schematic structural diagram of an electronic device provided by an embodiment of the present disclosure.
具体实施方式Detailed ways
现在将参考附图更全面地描述示例实施方式。然而，示例实施方式能够以多种形式实施，且不应被理解为限于在此阐述的范例；相反，提供这些实施方式使得本公开将更加全面和完整，并将示例实施方式的构思全面地传达给本领域的技术人员。所描述的特征、结构或特性可以以任何合适的方式结合在一个或更多实施方式中。Example embodiments will now be described more fully with reference to the accompanying drawings. Example embodiments may, however, be embodied in various forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the concepts of the example embodiments. To those skilled in the art. The described features, structures or characteristics may be combined in any suitable manner in one or more embodiments.
此外，附图仅为本公开的示意性图解，并非一定是按比例绘制。图中相同的附图标记表示相同或类似的部分，因而将省略对它们的重复描述。附图中所示的一些方框图是功能实体，不一定必须与物理或逻辑上独立的实体相对应。可以采用软件形式来实现这些功能实体，或在一个或多个硬件模块或集成电路中实现这些功能实体，或在不同网络和/或处理器装置和/或微控制器装置中实现这些功能实体。Furthermore, the drawings are merely schematic illustrations of the present disclosure and are not necessarily drawn to scale. The same reference numerals in the drawings represent the same or similar parts, and thus their repeated description will be omitted. Some of the block diagrams shown in the figures are functional entities and do not necessarily correspond to physically or logically separate entities. These functional entities may be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and/or processor devices and/or microcontroller devices.
如上所述，医学影像识别常采用卷积神经网络等深度学习算法。由于自身算法特点，其输出结果往往可解释性不佳。因此，医生很难将医学影像和算法输出结果联系起来，导致医生在面对输出结果时，难以从中得知其判断依据，进而妨碍其针对输出结果进行分析验证。As mentioned above, medical image recognition often uses deep learning algorithms such as convolutional neural networks. Due to its algorithm characteristics, its output results are often poorly interpretable. Therefore, it is difficult for doctors to connect medical images and algorithm output results, making it difficult for doctors to understand the basis for their judgment when faced with the output results, which further hinders their ability to analyze and verify the output results.
有鉴于此，本公开提供一种医学诊断装置，包括：第一获取模块，用于获取医学影像中的目标结构的第一异常特征信息，第一异常特征信息表征目标结构是否出现异常；第二获取模块，用于获取目标结构对应的疾病类型列表，疾病类型列表中包括发生于目标结构的多个疾病；诊断模块，用于基于第一异常特征信息和多个疾病之间的关联度，得到第一诊断信息。因此，本公开实施例提供的医学诊断装置能够通过医学影像所示出的目标结构中的异常特征，以及疾病与异常特征之间的关联关系，得到输出结果。其结果是基于疾病与异常特征之间的关联关系得到的，能够清晰地确定异常特征信息和疾病之间的对应关系，有着良好的可解释性。因此，医生能够根据本公开实施例的输出结果进行进一步地验证和定量分析，提高诊断效率和准确性，提高了临床医学影像的应用价值。In view of this, the present disclosure provides a medical diagnosis device, including: a first acquisition module for acquiring first abnormal feature information of a target structure in a medical image, where the first abnormal feature information represents whether the target structure is abnormal; second The acquisition module is used to obtain a list of disease types corresponding to the target structure. The disease type list includes multiple diseases that occur in the target structure; the diagnosis module is used to obtain based on the correlation between the first abnormal feature information and the multiple diseases. First diagnostic information. Therefore, the medical diagnosis device provided by embodiments of the present disclosure can obtain output results based on abnormal features in the target structure shown in the medical images and the correlation between the disease and the abnormal features. The results are based on the correlation between disease and abnormal features, which can clearly determine the correspondence between abnormal feature information and diseases, and have good interpretability. Therefore, doctors can conduct further verification and quantitative analysis based on the output results of embodiments of the present disclosure, thereby improving diagnostic efficiency and accuracy, and improving the application value of clinical medical images.
本公开实施例提供了一种医学诊断装置、电子设备及存储介质。该医学诊断装置具体可以集成在电子设备中，该电子设备可以是终端或服务器等设备。Embodiments of the present disclosure provide a medical diagnosis device, electronic equipment, and storage media. The medical diagnosis device can be integrated into an electronic device, and the electronic device can be a terminal or a server.
图1所示为本公开实施例所适用的一场景示意图。该场景包括医学诊断装置100，医学诊断装置100包括特征提取模块101和预测模块102。Figure 1 shows a schematic diagram of a scenario to which embodiments of the present disclosure are applicable. This scene includes a medical diagnosis device 100, which includes a feature extraction module 101 and a prediction module 102.
在一些实施例中，特征提取模块101用于从医学影像中提取患部特征；预测模块102用于对特征提取模块101从医学影像中提取的患部特征进行预测，得到诊断信息。In some embodiments, the feature extraction module 101 is used to extract affected part features from medical images; the prediction module 102 is used to predict the affected part features extracted from the medical images by the feature extraction module 101 to obtain diagnostic information.
在实际应用中，特征提取模块101能够从医学影像中提取医学影像所示目标结构的异常特征信息，并将异常特征信息传输至预测模块102中，预测模块102基于目标结构的异常特征信息进行预测，得到第一诊断信息，以辅助医生分析和判断医学影像。其中，医学影像可以是通过X射线成像、计算断层成像（CT）、核磁共振成像（MRI）、超声成像等方法拍摄的。In practical applications, the feature extraction module 101 can extract abnormal feature information of the target structure shown in the medical image from the medical image, and transmit the abnormal feature information to the prediction module 102. The prediction module 102 performs prediction based on the abnormal feature information of the target structure. , obtain the first diagnostic information to assist doctors in analyzing and judging medical images. Among them, medical images can be taken through X-ray imaging, computed tomography (CT), magnetic resonance imaging (MRI), ultrasound imaging and other methods.
在一些实施例中，医学诊断装置100还包括成像装置，成像装置用于获取相应的医学影像。根据医学影像所对应的目标结构不同，成像装置可以包括摄像机、X射线装置、电子计算机断层扫描装置或者核磁共振成像装置等。In some embodiments, the medical diagnostic device 100 further includes an imaging device, which is used to acquire corresponding medical images. Depending on the target structure corresponding to the medical image, the imaging device may include a camera, an X-ray device, a computerized tomography device, or an MRI device, etc.
下面将结合附图及实施例对本示例实施方式进行详细说明。This exemplary implementation will be described in detail below with reference to the drawings and examples.
图2示出本公开实施例中一种医学诊断装置的结构示意图，如图2所示，该医学诊断装置200包括：第一获取模块201、第二获取模块202和诊断模块203。Figure 2 shows a schematic structural diagram of a medical diagnosis device in an embodiment of the present disclosure. As shown in Figure 2, the medical diagnosis device 200 includes: a first acquisition module 201, a second acquisition module 202 and a diagnosis module 203.
第一获取模块201用于，获取医学影像中的目标结构的第一异常特征信息。The first acquisition module 201 is used to acquire first abnormal feature information of the target structure in the medical image.
目标结构为医学影像所示出的需要进行检查和诊断的部位。医学影像能够示出该目标结构的形态、尺寸、灰度信息、空间位置等特征，因此能够基于医学影像提取相应的第一异常特征信息，以表征目标结构是否出现异常，即当前目标结构的特征是否区别于目标结构的正常结构特征。The target structure is the area shown on the medical image that needs to be examined and diagnosed. Medical images can show the shape, size, grayscale information, spatial location and other characteristics of the target structure. Therefore, the corresponding first abnormal feature information can be extracted based on the medical image to characterize whether the target structure is abnormal, that is, the characteristics of the current target structure. Whether it is different from the normal structural characteristics of the target structure.
在一些实施例中，可由图像识别模型对医学影像进行数据预处理，清理数据，以提取其中的第一异常特征信息。其中，异常特征信息中包括多个分量，每个分量表征目标结构的一个结构特征是否出现异常，并以0、1进行二值编码。In some embodiments, the image recognition model can be used to perform data preprocessing on medical images and clean the data to extract the first abnormal feature information therein. Among them, the abnormal feature information includes multiple components, each component represents whether a structural feature of the target structure is abnormal, and is binary-coded with 0 and 1.
在一些实施例中，还可由医生对医学影像进行判断，以得到该医学影像示出的目标结构是否出现异常，并将判断结果输入医学诊断装置200，第一获取模块201生成相应的第一异常特征信息。In some embodiments, the doctor can also judge the medical image to determine whether the target structure shown in the medical image is abnormal, and input the judgment result into the medical diagnosis device 200, and the first acquisition module 201 generates the corresponding first abnormality. Feature information.
第二获取模块202用于，获取目标结构对应的疾病类型列表。The second obtaining module 202 is used to obtain a list of disease types corresponding to the target structure.
在一些实施例中，第二获取模块202能够根据目标结构，获取对应的疾病类型列表，疾病类型列表中包括发生于该目标结构的疾病类别，如炎症、癌变、组织损伤等。更具体地，疾病类型列表中还可以包括发生于该目标结构的多个多发的、有代表性的具体疾病名称。示例性地，若目标结构为中耳结构，那么对应的疾病类型列表中包括发生于中耳结构的多个疾病，例如，中耳炎、耳硬化症、鼓膜穿孔以及中耳肿瘤等疾病。若目标结构为肝脏结构，那么对应的疾病类型列表中包括发生于肝脏结构的多个疾病，例如，肝癌、脂肪肝、肝纤维化、药物性肝损伤、肝血管瘤等疾病。In some embodiments, the second acquisition module 202 can acquire a corresponding disease type list according to the target structure. The disease type list includes disease categories that occur in the target structure, such as inflammation, cancer, tissue damage, etc. More specifically, the disease type list may also include multiple common, representative and specific disease names that occur in the target structure. For example, if the target structure is a middle ear structure, the corresponding disease type list includes multiple diseases that occur in the middle ear structure, such as otitis media, otosclerosis, tympanic membrane perforation, and middle ear tumors. If the target structure is the liver structure, then the corresponding disease type list includes multiple diseases that occur in the liver structure, such as liver cancer, fatty liver, liver fibrosis, drug-induced liver injury, liver hemangioma and other diseases.
诊断模块203用于，基于第一异常特征信息和多个疾病之间的关联度，得到第一诊断信息。The diagnosis module 203 is configured to obtain first diagnosis information based on the correlation between the first abnormal feature information and multiple diseases.
在一些实施例中，第一诊断信息包括至少一个疾病，以及医学影像所示目标结构罹患该疾病的概率。其中，第一诊断信息中的疾病数量可以与疾病类型列表中的疾病数量相同，此时，第一诊断信息展示疾病类型列表中所有疾病的患病概率；或者，根据实际需要设置预设条件，使得第一诊断信息仅包括疾病对应的可能性高于预设阈值的疾病；或者，根据实际需要设置预设条件，使得第一诊断信息仅包括疾病对应的可能性排名前几位的疾病，以增强输出结果的灵活性，同时减少了部分计算量。In some embodiments, the first diagnostic information includes at least one disease and a probability that the target structure shown in the medical image suffers from the disease. The number of diseases in the first diagnosis information can be the same as the number of diseases in the disease type list. In this case, the first diagnosis information displays the disease probabilities of all diseases in the disease type list; or, set preset conditions according to actual needs, Make the first diagnostic information only include the diseases whose corresponding possibilities are higher than the preset threshold; or set the preset conditions according to actual needs, so that the first diagnostic information only includes the diseases whose corresponding possibilities are ranked among the top few, so as to Enhance the flexibility of output results while reducing some calculations.
在一些实施例中，由于目标结构形态病变与疾病之间存在关联关系，即，罹患某种疾病的目标结构，其形态的某些结构特征更可能区别于正常目标结构的结构特征。因此，根据该关联关系，能够量化第一异常特征信息与多个疾病之间的关联程度，得到第一异常特征信息和多个疾病之间的关联度。根据该关联度计算医学影像所示目标结构罹患疾病类型列表中各疾病的概率，并进一步地得到第一诊断信息。In some embodiments, due to the correlation between the morphological lesions of the target structure and the disease, that is, the target structure suffering from a certain disease may have certain structural characteristics of its morphology that are more likely to be different from the structural characteristics of the normal target structure. Therefore, based on the correlation relationship, the degree of correlation between the first abnormal characteristic information and multiple diseases can be quantified, and the degree of correlation between the first abnormal characteristic information and the multiple diseases can be obtained. According to the correlation degree, the probability that the target structure shown in the medical image suffers from each disease in the disease type list is calculated, and the first diagnostic information is further obtained.
因此，本实施例的装置能够获取医学影像中的目标结构的第一异常特征信息以及目标结构对应的疾病类型列表，根据第一异常特征信息和多个疾病之间的关联度，得到第一诊断信息。本实施例的输出结果是基于第一异常特征信息中各结构特征得到的，因此，根据本实施例装置的输出结果能够清晰地确定异常特征信息和疾病之间的对应关系，有着良好的可解释性；同时，医生在面对输出结果时，能够根据目标结构的结构特征，了解疾病的判断依据，以验证输出结果的可信度；此外，相关技术人员也可以根据医生验证结果进一步调整装置参数，进一步的提高装置输出结果的准确性。Therefore, the device of this embodiment can obtain the first abnormal feature information of the target structure in the medical image and the list of disease types corresponding to the target structure, and obtain the first diagnosis based on the correlation between the first abnormal feature information and multiple diseases. information. The output result of this embodiment is obtained based on each structural feature in the first abnormal feature information. Therefore, the output result of the device according to this embodiment can clearly determine the corresponding relationship between the abnormal feature information and the disease, and has good interpretability. At the same time, when faced with the output results, doctors can understand the basis for disease judgment based on the structural characteristics of the target structure to verify the credibility of the output results; in addition, relevant technical personnel can also further adjust the device parameters based on the doctor's verification results , further improving the accuracy of the device output results.
在一些实施例中，医学诊断装置应用于中耳病变诊断，第一异常特征信息由中耳图像提取。第一异常特征信息至少包括4个分量，分别为中耳形态信息、中耳图像灰度信息、中耳尺寸信息和中耳空间位置信息。中耳形态信息表征中耳形态是否出现异常，中耳图像灰度信息表征中耳图像灰度是否出现异常，中耳尺寸信息表征中耳尺寸是否出现异常，中耳空间位置信息表征中耳空间位置是否出现异常。示例性地，中耳形态信息可以为1或0，1表示中耳形态出现异常，0表示中耳形态未出现异常。In some embodiments, the medical diagnostic device is applied to the diagnosis of middle ear lesions, and the first abnormal feature information is extracted from the middle ear image. The first abnormal feature information includes at least four components, which are middle ear morphological information, middle ear image grayscale information, middle ear size information and middle ear spatial position information. The middle ear morphological information represents whether the middle ear morphology is abnormal, the middle ear image grayscale information represents whether the middle ear image grayscale is abnormal, the middle ear size information represents whether the middle ear size is abnormal, and the middle ear spatial position information represents the middle ear spatial position. Whether an exception occurs. For example, the middle ear morphology information may be 1 or 0. 1 indicates that the middle ear morphology is abnormal, and 0 indicates that there is no abnormality in the middle ear morphology.
在另一些实施例中，医学诊断装置应用于肝脏病变诊断，第一异常特征信息由肝脏图像提取。第一异常特征信息至少包括4个分量，分别为肝脏形态信息、肝脏图像灰度信息、肝脏尺寸信息和肝脏空间位置信息。肝脏形态信息表征肝脏形态是否出现异常，肝脏图像灰度信息表征肝脏图像灰度是否出现异常，肝脏尺寸信息表征肝脏尺寸是否出现异常，肝脏空间位置信息表征肝脏空间位置是否出现异常。示例性地，肝脏形态信息可以为1或0，1表示肝脏形态出现异常，0表示中耳形态未出现异常。可以理解的是，上述的第一异常特征信息所包括的分量仅是示意性的，第一异常特征信息中还可以包括其他的特征信息，具体可以根据实际应用场景确定。In other embodiments, the medical diagnostic device is applied to liver lesion diagnosis, and the first abnormal feature information is extracted from liver images. The first abnormal feature information includes at least four components, which are liver morphology information, liver image grayscale information, liver size information and liver spatial location information. The liver morphological information represents whether the liver morphology is abnormal, the liver image grayscale information represents whether the liver image grayscale is abnormal, the liver size information represents whether the liver size is abnormal, and the liver spatial position information represents whether the liver spatial position is abnormal. For example, the liver morphology information can be 1 or 0, with 1 indicating that the liver morphology is abnormal and 0 indicating that the middle ear morphology is not abnormal. It can be understood that the components included in the above-mentioned first abnormal characteristic information are only illustrative, and the first abnormal characteristic information may also include other characteristic information, and the specific information may be determined according to the actual application scenario.
在一些实施例中，诊断模块203用于基于第一异常特征信息和多个疾病之间的关联度，得到第一诊断信息；并且，如图3所示，基于第一异常特征信息和多个疾病之间的关联度，得到第一诊断信息，包括如下步骤。In some embodiments, the diagnosis module 203 is configured to obtain first diagnostic information based on the correlation between the first abnormal feature information and multiple diseases; and, as shown in Figure 3, based on the first abnormal feature information and multiple diseases The correlation between diseases and obtaining the first diagnostic information include the following steps.
S301，遍历疾病类型列表中的多个疾病。S301: Traverse multiple diseases in the disease type list.
具体地，遍历疾病类型列表中的多个疾病，分别基于第一异常特征信息与每个疾病进行如下计算，分别得到每个疾病对应的关联度。以此，能够确定与第一异常特征信息对应的每个疾病的可能性，避免遗漏，保证检查的全面性和完整性。Specifically, multiple diseases in the disease type list are traversed, and the following calculations are performed with each disease based on the first abnormal feature information to obtain the correlation degree corresponding to each disease. In this way, the possibility of each disease corresponding to the first abnormal characteristic information can be determined, omissions can be avoided, and the comprehensiveness and completeness of the examination can be ensured.
S302，根据目标结构对应的专家意见信息，计算第一异常特征信息与每个疾病之间的第一关联度。S302: Calculate the first degree of correlation between the first abnormal feature information and each disease based on the expert opinion information corresponding to the target structure.
具体地，可以通过收集专家意见，量化目标结构特征与疾病之间的关联程度，进而得到第一异常特征信息和多个疾病之间的第一关联度。Specifically, the degree of correlation between the target structural features and the disease can be quantified by collecting expert opinions, thereby obtaining the first degree of correlation between the first abnormal feature information and multiple diseases.
在一些实施例中，专家意见信息包括原始异常样本数据和与原始异常样本数据对应的专家诊断结果。其中，原始异常样本数据包括多张目标结构的图像，多张图像分别表现目标结构的多种结构特征正常和/或异常。并且，每张图像对应该图像所示目标结构可能罹患的疾病，即，专家诊断结果。每张图像可以对应于一个疾病；或者，每张图像可以对应于多个疾病。In some embodiments, the expert opinion information includes original abnormal sample data and expert diagnosis results corresponding to the original abnormal sample data. Among them, the original abnormal sample data includes multiple images of the target structure, and the multiple images respectively represent the normality and/or abnormality of various structural characteristics of the target structure. Moreover, each image corresponds to a disease that the target structure shown in the image may suffer from, that is, the expert diagnosis result. Each image can correspond to one disease; alternatively, each image can correspond to multiple diseases.
在一些实施例中，原始异常样本数据对应的专家诊断结果可以由德尔菲法（Delphi method）得到。德尔菲法是一种结构化的决策支持方法，采用函询调查的形式，依据系统的程序，采用匿名发表意见的方式，通过多轮次调查专家关于原始异常样本数据的看法，经过反复征询、归纳、修改和技术处理，最后汇总成专家基本一致的看法，作为最终的诊断结果。具体步骤如下。In some embodiments, the expert diagnosis results corresponding to the original abnormal sample data can be obtained by the Delphi method. The Delphi method is a structured decision support method. It adopts the form of correspondence survey, and adopts the method of anonymous expression of opinions according to systematic procedures. Through multiple rounds of investigation of experts' opinions on the original abnormal sample data, after repeated consultation and After induction, modification and technical processing, it is finally summarized into a basically unanimous opinion among experts as the final diagnostic result. Specific steps are as follows.
成立专家小组，该专家小组包括多位对应领域的医生或专家学者，并根据实际需要的知识范围及任务量，确定专家人数；Establish an expert group, which includes a number of doctors or experts and scholars in the corresponding field, and determine the number of experts based on the actual required knowledge scope and task volume;
向专家小组中的所有成员提供所要预测的问题及相关要求，并附上有关数据（即，原始异常样本数据）；Provide all members of the expert team with the problems to be predicted and related requirements, and attach relevant data (i.e., original anomaly sample data);
各成员根据收到的数据，给出自己的意见，包括数据是否属于异常，判断并标明是否正常或属于多类异常中的某类，以及一种或多种可能罹患的疾病；Each member gives his or her own opinion based on the data received, including whether the data is abnormal, judging and labeling whether it is normal or belongs to one of multiple types of abnormalities, and one or more possible diseases;
将第一次意见汇总，进行对比，将对比结果分送给各成员，让其根据对比结果修改自己的意见；或者，可以把各成员的意见加以整理，请专家小组成员之外的其他专家加以评论，再分送给各成员，以便他们参考后修改自己的意见；Summarize the first opinions, compare them, and distribute the comparison results to each member, allowing them to modify their opinions based on the comparison results; alternatively, you can organize the opinions of each member and ask other experts other than members of the expert group to review them. Comments are then sent to each member so that they can refer to them and modify their opinions;
将所有成员的意见收集起来，汇总，再次分送给各成员，以便做第二次修改。如此修改数轮，直到每个成员不再改变自己的意见为止。此时，将各成员的意见汇总，得到专家诊断结果。Collect the opinions of all members, summarize them, and distribute them to each member again for a second revision. Repeat this for several rounds until each member no longer changes his or her opinion. At this time, the opinions of each member are summarized and the expert diagnosis results are obtained.
在一些实施例中，如图4所示，根据目标结构对应的专家意见信息，计算第一异常特征信息与每个疾病之间的第一关联度进一步包括如下步骤。In some embodiments, as shown in Figure 4, calculating the first degree of correlation between the first abnormal feature information and each disease based on the expert opinion information corresponding to the target structure further includes the following steps.
S3021，基于原始异常样本数据，得到第二异常特征信息。S3021: Obtain the second abnormal feature information based on the original abnormal sample data.
具体地，可以基于图像识别方法或人工识别，从原始异常样本数据中提取第二异常特征信息。其中，第二异常特征信息表征原始异常样本数据中每张图像所示目标结构的多个结构特征是否出现异常。Specifically, the second abnormal feature information can be extracted from the original abnormal sample data based on the image recognition method or manual recognition. Among them, the second abnormal feature information represents whether multiple structural features of the target structure shown in each image in the original abnormal sample data are abnormal.
S3022，基于第二异常特征信息和专家诊断结果，得到疾病类型列表中每个疾病与患者体征之间的相关性。S3022: Based on the second abnormal feature information and expert diagnosis results, obtain the correlation between each disease in the disease type list and the patient's physical signs.
具体地，通过上述步骤，能够得到原始异常样本数据中每张图像各自对应的专家诊断结果，和第二异常特征信息。因此，能够通过统计计算，得到每个疾病和患者体征之间的对应关系，即相关性。Specifically, through the above steps, the expert diagnosis results corresponding to each image in the original abnormal sample data and the second abnormal feature information can be obtained. Therefore, the correspondence between each disease and the patient's signs, that is, the correlation, can be obtained through statistical calculations.
S3023，利用注意力机制，基于每个疾病与患者体征之间的相关性和第一异常特征信息，得到第一异常特征信息与每个疾病之间的第一关联度。S3023: Using the attention mechanism, based on the correlation between each disease and the patient's physical signs and the first abnormal feature information, obtain the first correlation between the first abnormal feature information and each disease.
具体地，注意力机制能够根据信息的重要程度，为其赋予不同的权重，使得计算资源分配给更重要的信息，降低对其他信息的关注度，从而提高输出结果的准确性和计算效率。Specifically, the attention mechanism can assign different weights to information according to its importance, so that computing resources are allocated to more important information and reduce attention to other information, thereby improving the accuracy and computing efficiency of the output results.
在一些实施例中，可以基于每个疾病与患者体征之间的相关性，将第二异常特征信息分组，每组对应于一种疾病；分别计算每组第二异常特征信息与第一异常特征信息之间的相似度（Similarity），并为相似度更大的第二异常特征信息赋予更高的权重；根据权重，将第二异常特征信息对应的专家诊断结果加权求和，最终得到第一异常特征信息与每个疾病之间的第一关联度。In some embodiments, the second abnormal feature information can be grouped based on the correlation between each disease and the patient's physical signs, with each group corresponding to one disease; each group of second abnormal feature information and the first abnormal feature are calculated separately Similarity between information (Similarity), and give a higher weight to the second abnormal feature information with greater similarity; according to the weight, the expert diagnosis results corresponding to the second abnormal feature information are weighted and summed, and finally the first The first degree of correlation between abnormal feature information and each disease.
本实施例中能够根据专家意见的客观表达，对第一异常特征信息进行分析判断，并引入注意力机制，提高输出结果的准确性和计算效率。因此，本实施例所得到的第一关联度基于专家意见信息预测，使得预测结果更贴近临床诊断结果。同时，该结果基于疾病与患者体征之间的相关性得到，因此有着良好的可解释性。In this embodiment, the first abnormal feature information can be analyzed and judged based on the objective expression of expert opinions, and an attention mechanism is introduced to improve the accuracy and calculation efficiency of the output results. Therefore, the first correlation obtained in this embodiment is predicted based on expert opinion information, making the prediction results closer to the clinical diagnosis results. At the same time, the results are based on the correlation between the disease and the patient's signs, so they have good interpretability.
S303，根据目标结构对应的历史诊断信息，计算第一异常特征信息与每个疾病之间的第二关联度。S303: Calculate the second degree of correlation between the first abnormal feature information and each disease based on the historical diagnosis information corresponding to the target structure.
具体地，可以根据目标结构对应的历史诊断信息，量化目标结构特征与疾病之间的关联程度，进而得到第一异常特征信息和多个疾病之间的第二关联度。Specifically, the degree of correlation between the characteristics of the target structure and the disease can be quantified based on the historical diagnosis information corresponding to the target structure, and then the second degree of correlation between the first abnormal characteristic information and multiple diseases can be obtained.
在一些实施例中，历史诊断信息包括第三异常特征信息和与第三异常特征信息对应的历史诊断结果。其中，第三异常特征信息和历史诊断结果由历史病例数据获取，历史病例数据包括目标结构的多张图像，以及每张图像对应的历史诊断结果；第三异常特征信息可以基于图像识别方法或人工识别从图像中提取，第三异常特征信息表征历史病例数据中每张图像所示目标结构的多个结构特征是否出现异常。In some embodiments, the historical diagnosis information includes third abnormal characteristic information and historical diagnosis results corresponding to the third abnormal characteristic information. Among them, the third abnormal feature information and historical diagnosis results are obtained from historical case data. The historical case data includes multiple images of the target structure and the historical diagnosis results corresponding to each image; the third abnormal feature information can be based on image recognition methods or artificial intelligence. The identification is extracted from the image, and the third abnormal feature information represents whether multiple structural features of the target structure shown in each image in the historical case data are abnormal.
在一些实施例中，如图5所示，根据目标结构对应的历史诊断信息，计算第一异常特征信息与每个疾病之间的第二关联度，包括如下步骤。In some embodiments, as shown in Figure 5, calculating the second degree of correlation between the first abnormal feature information and each disease based on the historical diagnosis information corresponding to the target structure includes the following steps.
S3031，计算第三异常特征信息和历史诊断结果间的频繁项集。S3031: Calculate frequent item sets between the third abnormal feature information and historical diagnosis results.
具体地，频繁项集是指数据集中经常一起出现的变量的集合。项集是指若干个项的集合；一个项集的支持度被定义为数据集中包含该项集的记录所占的比例；因此，可以将支持度不小于最小支持度的项集定义为频繁项集。因此，频繁项集能够表示数据集中经常一起出现的变量。其中，最小支持度可以根据实际需求人为预先设置，也可由某些算法计算得到。Specifically, a frequent itemset refers to a set of variables that often appear together in a data set. An item set refers to a collection of several items; the support of an item set is defined as the proportion of records in the data set that contains the item set; therefore, an item set whose support is not less than the minimum support can be defined as a frequent item set. Therefore, frequent itemsets can represent variables that often appear together in the data set. Among them, the minimum support can be preset manually according to actual needs, or it can be calculated by some algorithms.
S3032，基于频繁项集，得到疾病类型列表中每个疾病与患者体征之间的相关性。S3032: Based on the frequent item set, obtain the correlation between each disease in the disease type list and the patient's signs.
具体地，可以对频繁项集进行统计计算，或经某算法处理，得到疾病与患者体征之间的关联规则。关联规则的可信度表示该关联规则的可靠程度，示例性地，若关联规则为A→B，其可信度为：支持度({A, B})/支持度({A})，其中，A和B分别为一类项集。最后，基于可信度不小于最小可信度的关联规则，确定疾病类型列表中每个疾病与患者体征之间的相关性。其中，最小可信度可以根据实际需求人为预先设置，也可由某些算法计算得到。Specifically, frequent item sets can be statistically calculated or processed by an algorithm to obtain association rules between diseases and patient signs. The credibility of an association rule indicates the reliability of the association rule. For example, if the association rule is A→B, its credibility is: support ({A, B})/support ({A}), Among them, A and B are itemsets of one type respectively. Finally, the correlation between each disease in the disease type list and the patient's signs is determined based on association rules whose credibility is not less than the minimum credibility. Among them, the minimum credibility can be manually preset according to actual needs, or it can be calculated by some algorithms.
在一些实施例中，可由Apriori算法计算得到频繁项集，进而得到关联规则。首先，找出所有的频集，这些项集出现的频繁性至少和预定义的最小支持度相同；然后由频集产生强关联规则，这些规则必须满足最小支持度和最小可信度；然后使用上述频集产生期望的规则，产生只包含集合的项的所有规则；保留所有规则中可信度不小于最小可信度的规则。以此，利用递归的方法得到最终的关联规则。In some embodiments, frequent itemsets can be calculated by the Apriori algorithm, and then association rules can be obtained. First, find all frequency sets whose frequency of occurrence is at least the same as the predefined minimum support; then generate strong association rules from the frequency sets, and these rules must satisfy the minimum support and minimum credibility; then use The above frequency set generates the desired rules, generating all rules that only contain items of the set; retaining all rules whose credibility is not less than the minimum credibility. In this way, the final association rules are obtained using the recursive method.
S3033，利用注意力机制，基于每个疾病与患者体征之间的相关性和第一异常特征信息，得到第一异常特征信息与每个疾病之间的第二关联度。S3033: Using the attention mechanism, based on the correlation between each disease and the patient's physical signs and the first abnormal feature information, obtain the second degree of correlation between the first abnormal feature information and each disease.
如上所述，可以基于每个疾病与患者体征之间的相关性，将第三异常特征信息分组，每组对应于一种疾病；分别计算每组第三异常特征信息与第一异常特征信息之间的相似度，并为相似度更大的第三异常特征信息赋予更高的权重；根据权重，将第三异常特征信息对应的历史诊断结果加权求和，最终得到第一异常特征信息与每个疾病之间的第二关联度。As mentioned above, the third abnormal feature information can be grouped based on the correlation between each disease and the patient's physical signs, and each group corresponds to one disease; the difference between the third abnormal feature information and the first abnormal feature information of each group is calculated respectively. similarity between them, and give a higher weight to the third abnormal feature information with greater similarity; according to the weight, weight and sum the historical diagnosis results corresponding to the third abnormal feature information, and finally obtain the first abnormal feature information and each The second degree of correlation between diseases.
本实施例中能够根据历史数据的客观规律，对第一异常特征信息进行分析判断，并引入注意力机制，提高输出结果的准确性和计算效率。因此，本实施例所得到的第一关联度基于历史数据预测，使得预测结果更贴近统计结果。同时，该结果基于疾病与患者体征之间的相关性得到，因此有着良好的可解释性。In this embodiment, the first abnormal feature information can be analyzed and judged based on the objective rules of historical data, and an attention mechanism is introduced to improve the accuracy and calculation efficiency of the output results. Therefore, the first correlation degree obtained in this embodiment is based on historical data prediction, making the prediction result closer to the statistical result. At the same time, the results are based on the correlation between the disease and the patient's signs, so they have good interpretability.
S304，基于第一异常特征信息与每个疾病之间的第一关联度和第二关联度，得到第一诊断信息。S304: Obtain first diagnostic information based on the first degree of correlation and the second degree of correlation between the first abnormal characteristic information and each disease.
具体的，可以通过预设权重将第一关联度和第二关联度结合，并根据结合后的权重得到第一诊断信息。其中，预设权重可基于实验结果和/或专家意见与历史数据间的可靠程度确定。灵活地调整第一关联度和第二关联度间的权重，以提高输出结果的准确性。Specifically, the first degree of correlation and the second degree of correlation can be combined by preset weights, and the first diagnosis information can be obtained according to the combined weights. The preset weight may be determined based on the reliability between experimental results and/or expert opinions and historical data. Flexibly adjust the weight between the first correlation degree and the second correlation degree to improve the accuracy of the output results.
在一些实施例中，如图6所示，基于第一异常特征信息与每个疾病之间的第一关联度和第二关联度，得到第一诊断信息，包括如下步骤。In some embodiments, as shown in Figure 6, obtaining the first diagnostic information based on the first degree of correlation and the second degree of correlation between the first abnormal feature information and each disease includes the following steps.
S3041，基于第一异常特征信息与每个疾病之间的第一关联度，得到第二诊断信息。S3041: Obtain second diagnostic information based on the first correlation between the first abnormal characteristic information and each disease.
具体地，可以将每个疾病对应的第一关联度归一化，将归一化后的数据确定为每个疾病对应的概率。该某个疾病对应的概率表示，第一异常特征信息对应的图像被诊断为该疾病的概率。可以理解的是，第一异常特征信息分别对应多个疾病，以及罹患该疾病的多个概率。基于概率，可以进一步得到第二诊断信息。Specifically, the first correlation degree corresponding to each disease can be normalized, and the normalized data is determined as the probability corresponding to each disease. The probability corresponding to a certain disease represents the probability that the image corresponding to the first abnormal feature information is diagnosed with the disease. It can be understood that the first abnormal feature information respectively corresponds to multiple diseases and multiple probabilities of suffering from the diseases. Based on the probability, second diagnostic information can be further obtained.
在一些实施例中，基于第一异常特征信息与每个疾病之间的第一关联度，得到第二诊断信息，包括：将疾病类型列表中的多个疾病按照第一关联度由大到小的顺序排列，得到疾病队列；基于疾病队列，生成第二诊断信息。In some embodiments, obtaining the second diagnostic information based on the first degree of correlation between the first abnormal characteristic information and each disease includes: sorting multiple diseases in the disease type list from large to small according to the first degree of correlation. Arrange in order to obtain a disease queue; based on the disease queue, generate second diagnostic information.
示例性地，可以按照多个疾病所对应的第一关联度排序，或者，可以按照多个疾病所对应概率排序，以得到疾病队列。在得到疾病队列后，便可根据疾病队列确定第二诊断信息中所包括的疾病。For example, the diseases can be sorted according to the first correlation degree corresponding to the multiple diseases, or the diseases can be sorted according to the probabilities corresponding to the multiple diseases to obtain a disease queue. After the disease queue is obtained, the diseases included in the second diagnosis information can be determined according to the disease queue.
示例性地，可以根据实际需求，预先确定第二诊断信息中包括的疾病的数量，并选取疾病队列中前预设数量个疾病作为第二诊断信息。例如，若只想得到一个诊断结果，则将预设数量设置为一，从而选取疾病队列中排序为第一个的疾病；若想得到n个诊断结果，则将预设数量设置为n，从而选取疾病队列中前n个的疾病，以辅助医生根据多个信息综合地判断分析。For example, the number of diseases included in the second diagnosis information may be predetermined according to actual needs, and the first preset number of diseases in the disease queue may be selected as the second diagnosis information. For example, if you only want to get one diagnosis result, set the preset number to one to select the disease ranked first in the disease queue; if you want to get n diagnosis results, set the preset number to n to select the disease The top n diseases in the queue are used to assist doctors in comprehensive judgment and analysis based on multiple pieces of information.
通过这种设置，能够灵活地输出个性化的输出结果，提高使用体验。Through this setting, personalized output results can be flexibly output and the user experience can be improved.
S3042，基于第一异常特征信息与每个疾病之间的第二关联度，得到第三诊断信息。S3042: Obtain third diagnostic information based on the second degree of correlation between the first abnormal feature information and each disease.
具体地，可以将每个疾病对应的第二关联度归一化，将归一化后的数据确定为每个疾病对应的概率。基于概率，可以进一步得到第三诊断信息。Specifically, the second correlation degree corresponding to each disease can be normalized, and the normalized data is determined as the probability corresponding to each disease. Based on the probability, third diagnostic information can be further obtained.
S3043，基于第二诊断信息和第三诊断信息中每个疾病的概率，得到第一诊断信息。S3043: Obtain the first diagnostic information based on the probability of each disease in the second diagnostic information and the third diagnostic information.
具体地，可以根据第二诊断信息和第三诊断信息的可靠程度确定一预设权重，根据预设权重将第二诊断信息和第三诊断信息中每个疾病的概率加权求和。并根据最终的概率，将多个疾病排序，选取合适数量的疾病作为第一诊断信息。Specifically, a preset weight may be determined based on the reliability of the second diagnostic information and the third diagnostic information, and the probability of each disease in the second diagnostic information and the third diagnostic information may be weighted and summed based on the preset weight. And according to the final probability, multiple diseases are sorted, and an appropriate number of diseases are selected as the first diagnostic information.
通过这种设置，能够综合专家意见的客观表达以及历史数据的客观规律，从多种角度预测第一异常特征信息。专家意见依赖临床专家先验知识，临床主观经验影响比重大；同时，历史数据通过对历史数据的学习，从中获得一定规律，以此来预测可能的诊断结果，二者各有利弊。因此，本实施例通过专家意见和历史数据的可靠程度，灵活地结合二者的输出结果，相互弥补，得到最终的第一诊断信息。因此，能够提高输出结果的准确性和可信度。Through this setting, the objective expression of expert opinions and the objective laws of historical data can be combined to predict the first abnormal characteristic information from multiple perspectives. Expert opinions rely on the prior knowledge of clinical experts, and clinical subjective experience has a greater impact; at the same time, historical data can be used to predict possible diagnostic results by learning certain patterns from historical data. Both have their own advantages and disadvantages. Therefore, this embodiment uses the reliability of expert opinions and historical data to flexibly combine the output results of the two to complement each other and obtain the final first diagnostic information. Therefore, the accuracy and credibility of the output results can be improved.
此外，还可以通过一些机器学习算法对第一异常特征信息加以分析，进一步地提高输出结果的准确性。如图7所示，诊断模块203还用于实施以下步骤。In addition, some machine learning algorithms can also be used to analyze the first abnormal feature information to further improve the accuracy of the output results. As shown in Figure 7, the diagnostic module 203 is also used to implement the following steps.
S701，将第一异常特征信息输入预测模型，得到第四诊断信息。S701: Input the first abnormal feature information into the prediction model to obtain fourth diagnostic information.
在一些实施例中，可以将上述专家意见信息和/或历史诊断信息作为训练数据集，训练预测模型。预测模型能够基于输入的第一异常特征信息，预测其对应的第四诊断信息，其中，第四诊断信息包括至少一个疾病，以及第一异常特征信息所示目标结构罹患该疾病的概率。In some embodiments, the above expert opinion information and/or historical diagnosis information can be used as a training data set to train the prediction model. The prediction model can predict corresponding fourth diagnostic information based on the input first abnormal feature information, where the fourth diagnostic information includes at least one disease and the probability that the target structure indicated by the first abnormal feature information suffers from the disease.
在一些实施例中，预测模型可以是随机森林模型。随机森林模型能够在训练数据集中随机抽样，构成n个不同的样本数据集；根据n个不同的样本数据集，搭建n个不同的决策树；最后，根据n个决策树的投票结果，确定最终结果。In some embodiments, the predictive model may be a random forest model. The random forest model can randomly sample the training data set to form n different sample data sets; build n different decision trees based on n different sample data sets; finally, based on the voting results of the n decision trees, determine the final result.
通过随机森林模型，能够简单的处理高维度数据，无需进行特征选择，并能够有效地防止模型过拟合。并且，通过多决策树投票的方式，提高了输出结果的准确性。Through the random forest model, high-dimensional data can be easily processed without feature selection, and it can effectively prevent model overfitting. Moreover, the accuracy of the output results is improved through multi-decision tree voting.
S702，基于第一诊断信息和第四诊断信息，得到第五诊断信息。S702: Obtain fifth diagnostic information based on the first diagnostic information and the fourth diagnostic information.
上述预测模型虽然有着良好的准确性，但其预测过程是黑盒的，可解释性差。因此，可以将第四诊断信息和第一诊断信息有机的结合，得到第五诊断信息，进一步提高输出结果的准确性。此时，输出结果结合了两部分诊断信息的优势，有着良好的准确性和可信度，并且有着良好的可解释性。Although the above prediction model has good accuracy, its prediction process is black box and has poor interpretability. Therefore, the fourth diagnostic information and the first diagnostic information can be organically combined to obtain the fifth diagnostic information, further improving the accuracy of the output result. At this time, the output result combines the advantages of the two parts of diagnostic information, has good accuracy and credibility, and has good interpretability.
以上实施例中介绍了从医学影像中获取第一异常特征信息，并举例介绍了从不同目标结构对应的医学影像中获取第一异常特征信息的具体实施方式。在实际诊断中，检验人员针对目标结构的医学影像会出具相关的诊断报告。同时，该医学影像也可能关联患者的病史、体格检查、病理报告等文本信息，以及数值性检查报告中的数值信息。因此，可以进一步地从上述信息中提取目标结构的异常特征信息。The above embodiments introduce the acquisition of first abnormal feature information from medical images, and give examples of specific implementation methods for obtaining first abnormal feature information from medical images corresponding to different target structures. In actual diagnosis, the examiner will issue relevant diagnostic reports based on the medical images of the target structure. At the same time, the medical image may also be associated with text information such as the patient's medical history, physical examination, pathology report, and numerical information in the numerical examination report. Therefore, the abnormal feature information of the target structure can be further extracted from the above information.
具体地，如图8所示，获取医学影像中的目标结构的第一异常特征信息，包括如下步骤。Specifically, as shown in Figure 8, obtaining the first abnormal feature information of the target structure in the medical image includes the following steps.
S801，基于医学影像提取影像异常特征信息，影像异常特征信息包括：目标结构的形态信息、目标结构图像的灰度信息、目标结构的尺寸信息和目标结构的空间位置信息。S801. Extract image abnormal feature information based on medical images. The image abnormal feature information includes: morphological information of the target structure, grayscale information of the target structure image, size information of the target structure, and spatial position information of the target structure.
S802，获取与医学影像对应的病历文本，并基于病历文本提取文本异常特征信息。S802: Obtain medical record text corresponding to the medical image, and extract text abnormal feature information based on the medical record text.
病史信息可以包括患者的病史信息、体格检查信息、病理报告信息、医学影像的诊断报告等。文本异常特征信息可以包括：是否有与目标结构相关的既往病史、家族遗传史，是否抽烟、饮酒等信息。文本异常特征信息可以用1或0表示，1表示该项特征出现异常，0表示该项特征未出现异常。Medical history information may include patient medical history information, physical examination information, pathology report information, medical imaging diagnostic reports, etc. Text abnormal feature information can include: whether there is past medical history, family genetic history, smoking, drinking, etc. related to the target structure. Text anomaly feature information can be represented by 1 or 0. 1 indicates that the feature is abnormal, and 0 indicates that the feature is not abnormal.
S803，获取与医学影像对应的数值性检查报告，并基于数值性检查报告提取数值异常特征信息。S803: Obtain a numerical inspection report corresponding to the medical image, and extract numerical abnormality feature information based on the numerical inspection report.
数值性检查报告可以包括患者的激素、血常规、肿瘤标志物检查报告等数值性检查报告。数值异常特征信息可以包括：激素浓度是否异常（是否在参考范围内）、白细胞浓度是否异常、肿瘤标志物浓度是否异常等信息。数值异常特征信息可以用1或0表示，1表示该项特征出现异常，0表示该项特征未出现异常。Numerical examination reports may include numerical examination reports such as the patient's hormones, blood routine, and tumor marker examination reports. Numerical abnormality feature information may include: whether the hormone concentration is abnormal (whether it is within the reference range), whether the leukocyte concentration is abnormal, whether the tumor marker concentration is abnormal, and other information. Numerical anomaly feature information can be represented by 1 or 0. 1 indicates that the feature is abnormal, and 0 indicates that the feature is not abnormal.
S804，基于影像异常特征信息、文本异常特征信息、数值异常特征信息，确定第一异常特征信息。S804: Determine the first abnormal feature information based on the image abnormal feature information, text abnormal feature information, and numerical abnormal feature information.
由于目标结构的异常特征之间存在着关联关系，因此可以将异常特征组合成矩阵或者多维特征参数，以便得到多模态的第一异常特征信息。Since there is a correlation between abnormal features of the target structure, the abnormal features can be combined into a matrix or multi-dimensional feature parameters to obtain multi-modal first abnormal feature information.
从多方面获取到的目标结构的信息能够更准确的描述目标结构的具体情况，因此基于多模态的第一异常特征信息能够获得更准确的结果。示例性地，从医学影像中所提取的影像异常特征信息，以及从该医学影像的诊断报告中提取的文本异常特征信息之间显然是相关的，因此二者之间能够相互验证，进一步地提高了结果的准确性。The target structure information obtained from multiple aspects can more accurately describe the specific conditions of the target structure, so more accurate results can be obtained based on multi-modal first abnormal feature information. For example, the image abnormality feature information extracted from the medical image and the text abnormality feature information extracted from the diagnosis report of the medical image are obviously related, so the two can verify each other, further improving the accuracy of the results.
需要说明的是，上述实施例提供的医学诊断装置在用于医学诊断时，仅以上述各功能模块的划分进行举例说明，实际应用中，可以根据需要而将上述功能分配由不同的功能模块完成，即将装置的内部结构划分成不同的功能模块，以完成以上描述的全部或者部分功能。It should be noted that when the medical diagnosis device provided in the above embodiments is used for medical diagnosis, only the division of the above functional modules is used as an example. In practical applications, the above functions can be allocated to different functional modules according to needs. , that is, dividing the internal structure of the device into different functional modules to complete all or part of the functions described above.
所属技术领域的技术人员能够理解，本公开的各个方面可以实现为系统、方法或程序产品。因此，本公开的各个方面可以具体实现为以下形式，即：完全的硬件实施方式、完全的软件实施方式（包括固件、微代码等），或硬件和软件方面结合的实施方式，这里可以统称为“电路”、“模块”或“系统”。Those skilled in the art will understand that various aspects of the present disclosure may be implemented as systems, methods, or program products. Therefore, various aspects of the present disclosure may be embodied in the following forms, namely: a complete hardware implementation, a complete software implementation (including firmware, microcode, etc.), or an implementation combining hardware and software aspects, which may be collectively referred to herein as "Circuit", "Module" or "System".
基于同一发明构思，本公开实施例中还提供了一种电子设备900。图9显示的电子设备900仅仅是一个示例，不应对本公开实施例的功能和适用范围带来任何限制。Based on the same inventive concept, an electronic device 900 is also provided in an embodiment of the present disclosure. The electronic device 900 shown in FIG. 9 is only an example and should not impose any limitations on the functions and applicable scope of the embodiments of the present disclosure.
如图9所示，电子设备900以通用计算设备的形式表现。电子设备900的组件可以包括但不限于：上述至少一个处理单元910、上述至少一个存储单元920、连接不同系统组件（包括存储单元920和处理单元910）的总线930。As shown in Figure 9, electronic device 900 is embodied in the form of a general computing device. The components of the electronic device 900 may include, but are not limited to: the above-mentioned at least one processing unit 910, the above-mentioned at least one storage unit 920, and a bus 930 connecting different system components (including the storage unit 920 and the processing unit 910).
其中，存储单元存储有程序代码，程序代码可以被处理单元910执行，使得处理单元910执行本说明书中描述的根据本公开各种示例性实施方式的步骤。Wherein, the storage unit stores program code, and the program code can be executed by the processing unit 910, so that the processing unit 910 performs the steps described in this specification according to various exemplary embodiments of the present disclosure.
在一些实施例中，处理单元910可以执行一种医学诊断方法的如下步骤：获取医学影像中的目标结构的第一异常特征信息，所述第一异常特征信息表征所述目标结构是否出现异常；获取所述目标结构对应的疾病类型列表，所述疾病类型列表中包括发生于所述目标结构的多个疾病；基于所述第一异常特征信息和所述多个疾病之间的关联度，得到第一诊断信息。In some embodiments, the processing unit 910 can perform the following steps of a medical diagnosis method: obtain first abnormal feature information of a target structure in a medical image, where the first abnormal feature information represents whether the target structure is abnormal; Obtain a disease type list corresponding to the target structure, the disease type list includes multiple diseases that occur in the target structure; based on the correlation between the first abnormal feature information and the multiple diseases, obtain First diagnostic information.
存储单元920可以包括易失性存储单元形式的可读介质，例如随机存取存储单元（RAM）9201和/或高速缓存存储单元9202，还可以进一步包括只读存储单元（ROM）9203。The storage unit 920 may include a readable medium in the form of a volatile storage unit, such as a random access storage unit (RAM) 9201 and/or a cache storage unit 9202 , and may further include a read-only storage unit (ROM) 9203 .
存储单元920还可以包括具有一组（至少一个）程序模块9205的程序/实用工具9204，这样的程序模块9205包括但不限于：操作系统、一个或者多个应用程序、其它程序模块以及程序数据，这些示例中的每一个或某种组合中可能包括网络环境的实现。Storage unit 920 may also include a program/utility 9204 having a set of (at least one) program modules 9205 including, but not limited to: an operating system, one or more application programs, other program modules, and program data, Each of these examples, or some combination, may include the implementation of a network environment.
总线930可以为表示几类总线结构中的一种或多种，包括存储单元总线或者存储单元控制器、外围总线、图形加速端口、处理单元或者使用多种总线结构中的任意总线结构的局域总线。Bus 930 may be a local area representing one or more of several types of bus structures, including a memory unit bus or memory unit controller, a peripheral bus, a graphics acceleration port, a processing unit, or using any of a variety of bus structures. bus.
电子设备900也可以与一个或多个外部设备940（例如键盘、指向设备、蓝牙设备等）通信，还可与一个或者多个使得用户能与该电子设备900交互的设备通信，和/或与使得该电子设备900能与一个或多个其它计算设备进行通信的任何设备（例如路由器、调制解调器等等）通信。这种通信可以通过输入/输出（I/O）接口950进行。并且，电子设备900还可以通过网络适配器960与一个或者多个网络（例如局域网（LAN），广域网（WAN）和/或公共网络，例如因特网）通信。如图所示，网络适配器960通过总线930与电子设备900的其它模块通信。应当明白，尽管图中未示出，可以结合电子设备900使用其它硬件和/或软件模块，包括但不限于：微代码、设备驱动器、冗余处理单元、外部磁盘驱动阵列、RAID系统、磁带驱动器以及数据备份存储系统等。Electronic device 900 may also communicate with one or more external devices 940 (e.g., keyboard, pointing device, Bluetooth device, etc.), may also communicate with one or more devices that enable a user to interact with electronic device 900, and/or with Any device (eg, router, modem, etc.) that enables the electronic device 900 to communicate with one or more other computing devices. This communication may occur through input/output (I/O) interface 950. Furthermore, the electronic device 900 may also communicate with one or more networks (eg, a local area network (LAN), a wide area network (WAN), and/or a public network, such as the Internet) through the network adapter 960 . As shown, network adapter 960 communicates with other modules of electronic device 900 via bus 930. It should be understood that, although not shown in the figures, other hardware and/or software modules may be used in conjunction with electronic device 900, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives And data backup storage system, etc.
通过以上的实施方式的描述，本领域的技术人员易于理解，这里描述的示例实施方式可以通过软件实现，也可以通过软件结合必要的硬件的方式来实现。因此，根据本公开实施方式的技术方案可以以软件产品的形式体现出来，该软件产品可以存储在一个非易失性存储介质（可以是CD-ROM，U盘，移动硬盘等）中或网络上，包括若干指令以使得一台计算设备（可以是个人计算机、服务器、终端装置、或者网络设备等）执行根据本公开实施方式的方法。Through the above description of the embodiments, those skilled in the art can easily understand that the example embodiments described here can be implemented by software, or can be implemented by software combined with necessary hardware. Therefore, the technical solution according to the embodiment of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, U disk, mobile hard disk, etc.) or on the network , including several instructions to cause a computing device (which may be a personal computer, a server, a terminal device, a network device, etc.) to execute a method according to an embodiment of the present disclosure.
在本公开的示例性实施例中，还提供了一种计算机可读存储介质，该计算机可读存储介质可以是可读信号介质或者可读存储介质。其上存储有能够实现一种医学诊断方法的程序产品。在一些可能的实施方式中，本公开的各个方面还可以实现为一种程序产品的形式，其包括程序代码，当程序产品在终端设备上运行时，程序代码用于使终端设备执行本说明书上述方法实施例部分描述的根据本公开各种示例性实施方式的步骤。In an exemplary embodiment of the present disclosure, a computer-readable storage medium is also provided, and the computer-readable storage medium may be a readable signal medium or a readable storage medium. A program product capable of implementing a medical diagnosis method is stored thereon. In some possible implementations, various aspects of the present disclosure can also be implemented in the form of a program product, which includes program code. When the program product is run on a terminal device, the program code is used to cause the terminal device to execute the above described instructions. The Method Examples section describes steps according to various exemplary embodiments of the present disclosure.
本公开中的计算机可读存储介质的更具体的例子可以包括但不限于：具有一个或多个导线的电连接、便携式计算机磁盘、硬盘、随机访问存储器（RAM）、只读存储器（ROM）、可擦式可编程只读存储器（EPROM或闪存）、光纤、便携式紧凑磁盘只读存储器（CD-ROM）、光存储器件、磁存储器件、或者上述的任意合适的组合。More specific examples of computer readable storage media in this disclosure may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard drives, random access memory (RAM), read only memory (ROM), Erasable programmable read-only memory (EPROM or flash memory), fiber optics, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the above.
在本公开中，计算机可读存储介质可以包括在基带中或者作为载波一部分传播的数据信号，其中承载了可读程序代码。这种传播的数据信号可以采用多种形式，包括但不限于电磁信号、光信号或上述的任意合适的组合。可读信号介质还可以是可读存储介质以外的任何可读介质，该可读介质可以发送、传播或者传输用于由指令执行系统、装置或者器件使用或者与其结合使用的程序。In this disclosure, a computer-readable storage medium may include a data signal propagated in baseband or as part of a carrier wave carrying readable program code therein. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A readable signal medium may also be any readable medium other than a readable storage medium that can send, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device.
可选地，计算机可读存储介质上包含的程序代码可以用任何适当的介质传输，包括但不限于无线、有线、光缆、RF等等，或者上述的任意合适的组合。Alternatively, program code embodied on a computer-readable storage medium may be transmitted using any suitable medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination of the above.
在具体实施时，可以以一种或多种程序设计语言的任意组合来编写用于执行本公开操作的程序代码，程序设计语言包括面向对象的程序设计语言。程序代码可以完全地在用户计算设备上执行、部分地在用户设备上执行、作为一个独立的软件包执行、部分在用户计算设备上部分在远程计算设备上执行、或者完全在远程计算设备或服务器上执行。在涉及远程计算设备的情形中，远程计算设备可以通过任意种类的网络，包括局域网（LAN）或广域网（WAN），连接到用户计算设备，或者，可以连接到外部计算设备（例如利用因特网服务提供商来通过因特网连接）。When implemented, program code for performing operations of the present disclosure may be written in any combination of one or more programming languages, including object-oriented programming languages. The program code may execute entirely on the user's computing device, partly on the user's device, as a stand-alone software package, partly on the user's computing device and partly on a remote computing device, or entirely on the remote computing device or server execute on. In situations involving remote computing devices, the remote computing device may be connected to the user computing device through any kind of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device, such as provided by an Internet service. business to connect via the Internet).
应当注意，尽管在上文详细描述中提及了用于动作执行的设备的若干模块或者单元，但是这种划分并非强制性的。实际上，根据本公开的实施方式，上文描述的两个或更多模块或者单元的特征和功能可以在一个模块或者单元中具体化。反之，上文描述的一个模块或者单元的特征和功能可以进一步划分为由多个模块或者单元来具体化。It should be noted that although several modules or units of equipment for action execution are mentioned in the above detailed description, this division is not mandatory. In fact, according to embodiments of the present disclosure, the features and functions of two or more modules or units described above may be embodied in one module or unit. Conversely, the features and functions of one module or unit described above may be further divided into being embodied by multiple modules or units.
此外，尽管在附图中以特定顺序描述了本公开中方法的各个步骤，但是，这并非要求或者暗示必须按照该特定顺序来执行这些步骤，或是必须执行全部所示的步骤才能实现期望的结果。附加的或备选的，可以省略某些步骤，将多个步骤合并为一个步骤执行，以及/或者将一个步骤分解为多个步骤执行等。Furthermore, although various steps of the methods of the present disclosure are depicted in the drawings in a specific order, this does not require or imply that the steps must be performed in that specific order, or that all of the illustrated steps must be performed to achieve the desired results. result. Additionally or alternatively, certain steps may be omitted, multiple steps may be combined into one step for execution, and/or one step may be decomposed into multiple steps for execution, etc.
本领域技术人员在考虑说明书及实践这里公开的发明后，将容易想到本公开的其它实施方案。本公开旨在涵盖本公开的任何变型、用途或者适应性变化，这些变型、用途或者适应性变化遵循本公开的一般性原理并包括本公开未公开的本技术领域中的公知常识或惯用技术手段。说明书和实施例仅被视为示例性的，本公开的真正范围和精神由所附的权利要求指出。Other embodiments of the disclosure will be readily apparent to those skilled in the art from consideration of the specification and practice of the invention disclosed herein. The present disclosure is intended to cover any variations, uses, or adaptations of the disclosure that follow the general principles of the disclosure and include common common sense or customary technical means in the technical field that are not disclosed in the disclosure. . It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the disclosure being indicated by the following claims.
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311387445.5ACN117116472B (en) | 2023-10-25 | 2023-10-25 | Medical diagnostic apparatus, electronic device, and storage medium |
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311387445.5ACN117116472B (en) | 2023-10-25 | 2023-10-25 | Medical diagnostic apparatus, electronic device, and storage medium |
| Publication Number | Publication Date |
|---|---|
| CN117116472Atrue CN117116472A (en) | 2023-11-24 |
| CN117116472B CN117116472B (en) | 2024-10-11 |
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311387445.5AActiveCN117116472B (en) | 2023-10-25 | 2023-10-25 | Medical diagnostic apparatus, electronic device, and storage medium |
| Country | Link |
|---|---|
| CN (1) | CN117116472B (en) |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1280091A2 (en)* | 1996-07-12 | 2003-01-29 | First Opinion Corporation | Computerized medical diagnostic system utilizing list-based processing |
| US20100250275A1 (en)* | 2007-10-18 | 2010-09-30 | Canon Kabushiki Kaisha | Diagnosis support apparatus, method of controlling diagnosis support apparatus, and program therefor |
| US20160364862A1 (en)* | 2015-06-12 | 2016-12-15 | Merge Healthcare Incorporated | Methods and Systems for Performing Image Analytics Using Graphical Reporting Associated with Clinical Images |
| US20170061087A1 (en)* | 2014-05-12 | 2017-03-02 | Koninklijke Philips N.V. | Method and system for computer-aided patient stratification based on case difficulty |
| CN108565020A (en)* | 2018-04-25 | 2018-09-21 | 上海市肺科医院 | A kind of pneumoconiosis diagnostic expert system |
| CN110362690A (en)* | 2019-07-04 | 2019-10-22 | 北京爱医生智慧医疗科技有限公司 | A kind of medical knowledge map construction method and device |
| CN111816301A (en)* | 2020-07-07 | 2020-10-23 | 平安科技（深圳）有限公司 | Medical inquiry assisting method, device, electronic equipment and medium |
| WO2021081850A1 (en)* | 2019-10-30 | 2021-05-06 | 未艾医疗技术(深圳)有限公司 | Vrds 4d medical image-based spine disease recognition method, and related devices |
| CN113793305A (en)* | 2021-08-23 | 2021-12-14 | 上海派影医疗科技有限公司 | Pathological image classification and identification method and system integrating multiple information |
| CN113808738A (en)* | 2021-09-18 | 2021-12-17 | 安徽爱朋科技有限公司 | Disease identification system based on self-identification image |
| CN114141358A (en)* | 2021-11-29 | 2022-03-04 | 平安医疗健康管理股份有限公司 | Disease diagnosis apparatus based on knowledge map, computer device, and storage medium |
| CN114446430A (en)* | 2022-01-28 | 2022-05-06 | 林星山 | Disease screening system based on remote images and outpatient service data |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1280091A2 (en)* | 1996-07-12 | 2003-01-29 | First Opinion Corporation | Computerized medical diagnostic system utilizing list-based processing |
| US20100250275A1 (en)* | 2007-10-18 | 2010-09-30 | Canon Kabushiki Kaisha | Diagnosis support apparatus, method of controlling diagnosis support apparatus, and program therefor |
| US20170061087A1 (en)* | 2014-05-12 | 2017-03-02 | Koninklijke Philips N.V. | Method and system for computer-aided patient stratification based on case difficulty |
| US20160364862A1 (en)* | 2015-06-12 | 2016-12-15 | Merge Healthcare Incorporated | Methods and Systems for Performing Image Analytics Using Graphical Reporting Associated with Clinical Images |
| CN108565020A (en)* | 2018-04-25 | 2018-09-21 | 上海市肺科医院 | A kind of pneumoconiosis diagnostic expert system |
| CN110362690A (en)* | 2019-07-04 | 2019-10-22 | 北京爱医生智慧医疗科技有限公司 | A kind of medical knowledge map construction method and device |
| WO2021081850A1 (en)* | 2019-10-30 | 2021-05-06 | 未艾医疗技术(深圳)有限公司 | Vrds 4d medical image-based spine disease recognition method, and related devices |
| CN111816301A (en)* | 2020-07-07 | 2020-10-23 | 平安科技（深圳）有限公司 | Medical inquiry assisting method, device, electronic equipment and medium |
| CN113793305A (en)* | 2021-08-23 | 2021-12-14 | 上海派影医疗科技有限公司 | Pathological image classification and identification method and system integrating multiple information |
| CN113808738A (en)* | 2021-09-18 | 2021-12-17 | 安徽爱朋科技有限公司 | Disease identification system based on self-identification image |
| CN114141358A (en)* | 2021-11-29 | 2022-03-04 | 平安医疗健康管理股份有限公司 | Disease diagnosis apparatus based on knowledge map, computer device, and storage medium |
| CN114446430A (en)* | 2022-01-28 | 2022-05-06 | 林星山 | Disease screening system based on remote images and outpatient service data |
| Publication number | Publication date |
|---|---|
| CN117116472B (en) | 2024-10-11 |
| Publication | Publication Date | Title |
|---|---|---|
| Bharati et al. | A review on explainable artificial intelligence for healthcare: Why, how, and when? | |
| Kaur et al. | Medical diagnostic systems using artificial intelligence (AI) algorithms: principles and perspectives | |
| Dack et al. | Artificial intelligence and interstitial lung disease: diagnosis and prognosis | |
| US8682693B2 (en) | Patient data mining for lung cancer screening | |
| US7711404B2 (en) | Patient data mining for lung cancer screening | |
| US20220051771A1 (en) | Report generating system and methods for use therewith | |
| Bozkurt et al. | Using automatically extracted information from mammography reports for decision-support | |
| Salih et al. | A review of evaluation approaches for explainable AI with applications in cardiology | |
| JP2018067303A (en) | Diagnosis support method, program and apparatus | |
| BRPI1006388A2 (en) | clinical decision support system (adc) and clinical decision support method (adc) implemented by an adc system | |
| JP2014505950A (en) | Imaging protocol updates and / or recommenders | |
| CN118039125A (en) | Pathological picture auxiliary diagnosis method and system based on artificial intelligence | |
| CN120221055A (en) | Clinical performance verification system based on infectious lung disease diagnostic model | |
| CN120148720A (en) | Rapid archiving method and system based on semantic extraction | |
| CN118522469B (en) | Big data analysis medical decision method and system | |
| CN118522428A (en) | Matching method, system, equipment and storage medium for remote medical resources | |
| CN117116472B (en) | Medical diagnostic apparatus, electronic device, and storage medium | |
| Sharma et al. | Intelligent Medical Diagnosis Model Based on Graph Neural Networks for Medical Images | |
| CN114550930A (en) | Disease prediction method, device, equipment and storage medium | |
| Farha et al. | Transforming pulmonary health care: the role of artificial intelligence in diagnosis and treatment | |
| CN120511020B (en) | An intelligent triage assistance system to improve the work efficiency of outpatient doctors in hospitals | |
| AU2021102832A4 (en) | System & method for automatic health prediction using fuzzy based machine learning | |
| Prasudha et al. | Research reviews: towards identification and classification kidney disease using computational technology | |
| Darmawahyuni et al. | Analysis of Classifier Performance on ECG Interpretation for Precision Medicine: Which performance metrics should we use? | |
| Chen et al. | A primer on artificial intelligence combined with medical images in sports |
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |