CN110946616A - Ultrasonic detection system and ultrasonic detection method based on cloud platform - Google Patents
Ultrasonic detection system and ultrasonic detection method based on cloud platformDownload PDFInfo
- Publication number
- CN110946616A CN110946616ACN201911319654.XACN201911319654ACN110946616ACN 110946616 ACN110946616 ACN 110946616ACN 201911319654 ACN201911319654 ACN 201911319654ACN 110946616 ACN110946616 ACN 110946616A
- Authority
- CN
- China
- Prior art keywords
- ultrasonic
- image
- cloud platform
- data file
- module
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000001514detection methodMethods0.000titleclaimsabstractdescription102
- 238000012545processingMethods0.000claimsabstractdescription36
- 230000011218segmentationEffects0.000claimsabstractdescription12
- 239000000284extractSubstances0.000claimsabstractdescription7
- 238000002604ultrasonographyMethods0.000claimsdescription65
- 238000013145classification modelMethods0.000claimsdescription15
- 238000012549trainingMethods0.000claimsdescription12
- 238000004806packaging method and processMethods0.000claimsdescription7
- 238000006243chemical reactionMethods0.000claimsdescription4
- 238000007689inspectionMethods0.000claimsdescription4
- 230000007175bidirectional communicationEffects0.000claimsdescription2
- 238000012360testing methodMethods0.000claimsdescription2
- 210000001685thyroid glandAnatomy0.000description13
- 238000000034methodMethods0.000description6
- 238000012216screeningMethods0.000description4
- 206010018498GoitreDiseases0.000description3
- 238000013473artificial intelligenceMethods0.000description3
- 230000017531blood circulationEffects0.000description3
- 238000003745diagnosisMethods0.000description3
- 201000003872goiterDiseases0.000description3
- 230000003902lesionEffects0.000description3
- 238000013507mappingMethods0.000description3
- 230000009286beneficial effectEffects0.000description2
- 230000000739chaotic effectEffects0.000description2
- 230000006854communicationEffects0.000description2
- 238000004891communicationMethods0.000description2
- 238000010586diagramMethods0.000description2
- 239000003814drugSubstances0.000description2
- 238000003384imaging methodMethods0.000description2
- 210000004185liverAnatomy0.000description2
- 238000003062neural network modelMethods0.000description2
- 239000000523sampleSubstances0.000description2
- 206010020850HyperthyroidismDiseases0.000description1
- 206010056342Pulmonary massDiseases0.000description1
- 208000009453Thyroid NoduleDiseases0.000description1
- 210000001367arteryAnatomy0.000description1
- 238000013528artificial neural networkMethods0.000description1
- 230000003247decreasing effectEffects0.000description1
- 238000005516engineering processMethods0.000description1
- 238000000605extractionMethods0.000description1
- 210000004072lungAnatomy0.000description1
- 208000024891symptomDiseases0.000description1
Images
Classifications
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B8/00—Diagnosis using ultrasonic, sonic or infrasonic waves
- A61B8/56—Details of data transmission or power supply
- A61B8/565—Details of data transmission or power supply involving data transmission via a network
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B8/00—Diagnosis using ultrasonic, sonic or infrasonic waves
- A61B8/52—Devices using data or image processing specially adapted for diagnosis using ultrasonic, sonic or infrasonic waves
- A61B8/5215—Devices using data or image processing specially adapted for diagnosis using ultrasonic, sonic or infrasonic waves involving processing of medical diagnostic data
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Heart & Thoracic Surgery (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Pathology (AREA)
- Radiology & Medical Imaging (AREA)
- Biomedical Technology (AREA)
- Veterinary Medicine (AREA)
- Medical Informatics (AREA)
- Physics & Mathematics (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- General Health & Medical Sciences (AREA)
- Public Health (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computer Networks & Wireless Communication (AREA)
- Ultra Sonic Daignosis Equipment (AREA)
Abstract
Description
Technical Field
The invention relates to the technical field of ultrasonic detection, in particular to an ultrasonic detection system and an ultrasonic detection method based on a cloud platform.
Background
At present, an Artificial Intelligence (AI) technology has been incorporated into an ultrasonic detection system, and applications of "AI + cloud platform + medical treatment" are all focus screening items focusing on medical images, such as lung nodules, thyroid nodules and the like. However, the current ultrasonic detection system has the following problems: (1) the application scenario of lesion screening is single and cannot cope with multiple application scenarios, for example, an ultrasonic detection system for detecting thyroid gland can only be used for detecting thyroid gland, but cannot be used for detecting lung; (2) the image data volume required by AI screening is relatively large, so the requirement on the computer hardware for processing is relatively high; (3) the screened and marked focus needs a doctor to manually select a report of a corresponding type to fill in, so that the doctor is required to be familiar with the ultrasonic workstation reporting system, and the doctor does not have much time to learn the classification of the entries of the ultrasonic workstation reporting system in the actual situation, so that in the actual use process, the doctor mostly uses a 'self-defining' function, and the function of the entries is greatly wasted.
Therefore, the invention provides an ultrasonic detection system and an ultrasonic detection method based on a cloud platform, wherein an image recognition module is preset on the cloud platform, and the image recognition model is trained through focus image characteristics prestored in the cloud platform to obtain a plurality of sectional recognition models so as to adapt to focus screening of different application scenes; and then, performing feature extraction on the ultrasonic image uploaded by the ultrasonic workstation by using a segmental recognition model to obtain key words corresponding to the focus of the current detection part, and searching out related entries through the key words to automatically generate an ultrasonic report so as to reduce the workload of detection personnel.
Disclosure of Invention
Aiming at the problems in the prior art, an ultrasonic detection system and an ultrasonic detection method based on a cloud platform are provided.
The specific technical scheme is as follows:
the invention comprises an ultrasonic detection system based on a cloud platform, which comprises an ultrasonic workstation and a cloud platform, wherein the ultrasonic workstation is in bidirectional communication connection with the cloud platform, and comprises:
an entry module for entering associated information of the patient, the associated information including detection position information indicating a current detection site;
the acquisition module is used for acquiring an ultrasonic image of the detection part;
the processing module is respectively connected with the input module and the acquisition module and is used for packaging the associated information and the ultrasonic image into an ultrasonic data file and then sending the ultrasonic data file to the cloud platform;
the cloud platform includes:
a semantic recognition module, which adopts a semantic recognition model formed by pre-training to perform semantic segmentation and recognition on the associated information in the ultrasonic data file, extracts the detection position information and generates a first class of keywords corresponding to the detection part;
the image recognition module is used for extracting image characteristics of the ultrasonic image in the ultrasonic data file by adopting an image recognition model formed by pre-training and generating a corresponding second class keyword according to the extracted image characteristics;
the entry retrieval module is respectively connected with the semantic recognition module and the image recognition module and is used for fusing the first class of keywords and the second class of keywords to form a core keyword and searching corresponding ultrasonic visible entries and ultrasonic suggested entries in an expert word library preset by the cloud platform according to the core keyword;
and the report generating module is connected with the entry retrieval module and used for generating at least one primary processing report according to the ultrasonic visible entry and the ultrasonic suggested entry obtained by searching and feeding the primary processing report back to the ultrasonic workstation.
Preferably, the processing module comprises:
the file generating unit is used for packaging the associated information and the ultrasonic image to form the ultrasonic data file and storing the associated information in a file header of the ultrasonic data file;
and the format conversion unit is connected with the file generation unit, converts the ultrasonic data file from a first image format to a second image format, and sends the ultrasonic data file converted into the second image format to the cloud platform.
Preferably, the ultrasound workstation further comprises:
and the first encryption module is connected with the processing module and used for encrypting the processed ultrasonic data file and sending the encrypted ultrasonic data file to the cloud platform.
Preferably, the cloud platform further comprises:
and the decryption module is used for decrypting the ultrasonic data file uploaded by the ultrasonic workstation.
Preferably, the cloud platform further comprises:
and the second encryption module is used for packaging the ultrasonic data file uploaded by the ultrasonic workstation and the primary processing report generated by the cloud platform, carrying out second encryption on the packaged ultrasonic data file and the primary processing report, and storing the encrypted ultrasonic data file and the encrypted primary processing report in the cloud platform.
Preferably, the image recognition model includes an image classification model and a plurality of segmentation recognition models, an input end of the image classification model is used as an input end of the image recognition model, an input end of each segmentation recognition model is respectively connected with an output end of the image classification model, and output ends of all the segmentation recognition models form an output end of the image recognition model;
the image classification model is used for extracting corresponding image characteristics according to the input ultrasonic image, classifying the detection part of the ultrasonic image according to the image characteristics and outputting a classification result;
each subsection identification model corresponds to one detection part, the image classification model inputs the classification result into the corresponding subsection identification model, and the second class of keywords are obtained through the identification of the subsection identification model and output;
the cloud platform further comprises:
and the training module is connected with the image recognition module and trains the image recognition model through different kinds of focus ultrasonic images prestored in the cloud platform.
Preferably, the first image format is a JPG format or a BMP format.
Preferably, the second image format is a DICOM format.
The invention comprises an ultrasonic detection method based on a cloud platform, which is applied to the ultrasonic detection system and comprises the following steps:
step S1, the detection personnel inputs the associated information of the patient into the ultrasonic workstation, and the associated information comprises the current detection part;
step S2, the inspector uses the ultrasonic workstation to acquire the current ultrasonic image of the detected part;
step S3, the ultrasound workstation packages the associated information and the ultrasound image into the ultrasound data file, stores the associated information in a file header of the ultrasound data file, and then sends the ultrasound data file to the cloud platform;
step S4, after receiving the ultrasonic data file, the cloud platform performs semantic segmentation on the associated information in the ultrasonic data file by using the semantic recognition model to extract the detection part, so as to generate the first class keywords corresponding to the detection part;
step S5, the cloud platform adopts the image recognition model to extract image features of the ultrasonic image in the ultrasonic data file, so as to generate the corresponding second type keywords according to the extracted image features;
step S6, the cloud platform fuses the first category of keywords and the second category of keywords to form the core keywords, and searches the corresponding ultrasonic visible entries and the corresponding ultrasonic suggested entries in an expert word library preset by the cloud platform according to the core keywords;
step S7, the cloud platform generates at least one preliminary treatment report by the ultrasonic visible term and the ultrasonic suggested term, and pushes the preliminary treatment report to the ultrasonic workstation.
The technical scheme of the invention has the beneficial effects that: the invention provides an ultrasonic detection system and an ultrasonic detection method based on a cloud platform, wherein the cloud platform respectively adopts a semantic recognition model and an image recognition model to extract first-class keywords and second-class keywords, then the first-class keywords and the second-class keywords are fused to form core keywords, and corresponding ultrasonic visible entries and ultrasonic suggested entries are retrieved through the core keywords, so that a preliminary processing report is automatically generated, the workload of detection personnel is reduced, and the detection efficiency of the detection personnel is improved.
Drawings
Embodiments of the present invention will be described more fully with reference to the accompanying drawings. The drawings are, however, to be regarded as illustrative and explanatory only and are not restrictive of the scope of the invention.
FIG. 1 is a schematic diagram of an ultrasonic inspection system in an embodiment of the invention;
fig. 2 is a schematic structural diagram of a DICOM image according to an embodiment of the present invention;
fig. 3 is a flowchart illustrating steps of an ultrasonic testing method according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
It should be noted that the embodiments and features of the embodiments may be combined with each other without conflict.
The invention is further described with reference to the following drawings and specific examples, which are not intended to be limiting.
The invention comprises an ultrasonic detection system based on a cloud platform, as shown in fig. 1, comprising anultrasonic workstation 1 and acloud platform 2, wherein theultrasonic workstation 1 comprises:
anentry module 101, configured to enter associated information of a patient, where the associated information includes detection position information indicating a current detection location;
anacquisition module 102, configured to acquire an ultrasound image of a detected portion;
theprocessing module 103 is respectively connected with therecording module 101 and theacquisition module 102, and is used for packaging the associated information and the ultrasonic image into an ultrasonic data file and then sending the ultrasonic data file to thecloud platform 2;
thecloud platform 2 includes:
asemantic recognition module 201, which performs semantic segmentation and recognition on the associated information in the ultrasonic data file by adopting a semantic recognition model formed by pre-training, extracts the detection position information, and generates a first class of keywords corresponding to the detection part;
animage recognition module 202, which extracts image features of an ultrasound image in an ultrasound data file by using an image recognition model formed by pre-training, and generates a corresponding second class keyword according to the extracted image features;
theentry retrieval module 203 is respectively connected with thesemantic recognition module 201 and theimage recognition module 202 and is used for fusing the first class of keywords and the second class of keywords to form a core keyword and searching corresponding ultrasonic visible entries and ultrasonic suggested entries in an expert word library preset by the cloud platform according to the core keyword;
and thereport generating module 204 is connected with theentry retrieving module 203 and is used for generating at least one preliminary processing report according to the ultrasonic found entries and the ultrasonic suggested entries obtained by searching and feeding back the preliminary processing report to theultrasonic workstation 1.
Specifically, in this embodiment, the present invention provides an ultrasonic detection system based on a cloud platform, where a processing procedure of the ultrasonic detection system includes: before ultrasonic detection is carried out, a detector inputs relevant information of a patient, wherein the relevant information comprises information of a detection part, detection time and the like related to the current detection of the patient and also comprises information of the name, age, sex, information of first and second rounds, a historical diagnosis report and the like of the patient; after the associated information is input, acquiring an ultrasonic image through an acquisition module, wherein the acquisition module is an ultrasonic probe, a tester places the ultrasonic probe on a detection part of a patient, a display of an ultrasonic workstation displays the ultrasonic image of the current detection part, and when a proper ultrasonic image is detected, the acquisition module acquires an ultrasonic image in a JPG or BMP format and stores the ultrasonic image in a local disk of the ultrasonic workstation; the processing module converts the acquired ultrasound image in the JPG/BMP format and the associated information of the corresponding patient together according to a preset processing rule to form an ultrasound data file, stores the associated information in a file header of the ultrasound data file according to the preset rule, stores the processed ultrasound data file to a local disk of an ultrasound workstation and uploads the processed ultrasound data file to a medical database of a cloud platform.
Further, after the cloud platform receives the ultrasound data file, the semantic recognition unit calls the ultrasound data file from the medical database, performs semantic segmentation on the associated information in the file header by using a semantic recognition model, and extracts the detection part information of the current ultrasound detection, so as to match a first type of keyword corresponding to the current detection part, for example, if the current detection part is a thyroid, the first type of keyword is the thyroid; then, the image recognition unit recognizes an ultrasonic image in an ultrasonic data file through an image recognition model, a plurality of segmental recognition models are preset in the image recognition model, each segmental model corresponds to a detection part, if the current detection part is identified to be a thyroid gland by the image recognition model, the ultrasonic image is input into the segmental recognition model corresponding to the thyroid gland for further recognition, so as to extract a second class of keywords corresponding to the ultrasonic image, for example, the extracted second class of keywords can be 'goiter', 'thyroid gland blood flow signal is rich' and the like, finally, the cloud platform fuses the first class of keywords and the second class of keywords, selects a core keyword which can most reflect the characteristics of the current focus, for example, 'goiter', and searches corresponding ultrasonic visible entries and ultrasonic suggested entries according to the core keyword 'goiter', the term of ultrasound is the content displayed by the ultrasound image, for example, the term of ultrasound related to thyroid has the symptoms of diffuse thyroid blood flow increase, blood flow velocity increase of superior thyroid artery, etc., and the term of ultrasound suggestion is the diagnosis suggestion corresponding to the term of ultrasound, for example, the term of ultrasound suggestion for thyroid enlargement can be the term of 'primarily considering hyperthyroidism, suggesting to visit the endocrinology department, and treating after definite diagnosis'. And finally, the cloud platform automatically generates 5 parts of preliminary processing reports according to the ultrasonic seen entries, the ultrasonic suggested entries and the ultrasonic images, stores the preliminary processing reports in the cloud platform, and simultaneously feeds the 5 parts of preliminary processing reports back to the ultrasonic workstation for reference of detection personnel, so that the working efficiency of the detection personnel is effectively improved.
In a preferred embodiment, as shown in FIG. 1, the processing module comprises:
a file generating unit 1031, which packs the associated information and the ultrasound image to form an ultrasound data file, and stores the associated information in a file header of the ultrasound data file;
a format conversion unit 1032 connected to the file generating unit 1031, converts the ultrasound data file from a first image format to a second image format, and sends the ultrasound data file converted to the second image format to thecloud platform 2.
Specifically, when the file generation unit names the ultrasound data files, the related information of the patients, such as the name, sex, age, time of image acquisition, etc., of the patients are concatenated as the file name of the ultrasound data files, and finally the labeled ultrasound data files are stored in a jpg/bmp/dcm format as a suffix, such as "zhangsan _ M19910101_201912051641_.
Specifically, the initial ultrasound image acquired by the ultrasound workstation is in a JPG format or a BMP format, and the format conversion unit combines the associated information of the patient with the ultrasound image in the JPG/BMP format, converts the associated information into a dicom (digital Imaging and communications in medicine) image, and uploads the dicom (digital Imaging and communications in medicine) image to the cloud platform. DICOM is an international standard for medical images and related information (ISO 12052) that defines a medical image format available for data exchange with quality that meets clinical needs. The DICOM image generally comprises a DICOM file header and a DICOM data set, wherein the DICOM data set is formed by arranging DICOM data elements according to a certain sequence, and comprises ultrasound image data and patient related information, such as patient name, age, medical history and the like. The data element is structured as shown in fig. 2, wherein the tag is composed of a group number and an element number, and the tag is a unique identifier of the data element.
In a preferred embodiment, as shown in fig. 1, theultrasound workstation 1 further comprises:
thefirst encryption module 104 is connected with theprocessing module 103 and used for encrypting the processed ultrasonic data file and sending the encrypted ultrasonic data file to thecloud platform 2;
specifically, in the present embodiment, the DICOM image is encrypted by using the improved AES encryption algorithm and then uploaded to thecloud platform 2. The encryption algorithm adopted by thefirst encryption module 104 is an improved AES algorithm, which is a block iterative encryption algorithm, can effectively resist strong attack, differential attack and linear cryptanalysis, and has the advantages of flexibility in relation to block length and key length, high security, high operating efficiency and the like. Here, an improved AES encryption algorithm is used, specifically as follows:
assuming that the DICOM image is an image with a size of MxN, 2 key sequences generated by using the following oblique tent mapping are respectively a column vector Km and a row vector Kn:
when a belongs to [0,1], the encryption system is in a chaotic state, the correlation of the mapping iteration track sequence is decreased by an index, the distribution of chaotic variables is uniform, and the pseudo-random characteristic is good.
By the encryption method, protection of the associated information of the patient is combined with encryption of ultrasonic image data, a one-dimensional to two-dimensional mapping method is designed, the associated information of the patient is exchanged with the encrypted ultrasonic image pixels by taking 8 bits as a unit, safety of an ultrasonic data file uploaded by an ultrasonic workstation can be ensured, and the associated information, detection information and the like of the patient are prevented from being leaked.
As a preferred embodiment, thecloud platform 2 further includes:
adecryption module 205, configured to decrypt the ultrasound data file uploaded by theultrasound workstation 1;
and thesecond encryption module 206 is configured to package the ultrasound data file uploaded by theultrasound workstation 1 and the preliminary processing report generated by thecloud platform 2, perform second encryption, and store the encrypted data file in thecloud platform 2.
Specifically, since the ultrasound data file uploaded by theultrasound workstation 1 is an encrypted file, after the ultrasound data file is uploaded to the cloud platform, the content of the ultrasound data file is first decrypted and then identified. After each ultrasonic detection, thecloud platform 2 packages the ultrasonic data file detected this time and the finally formed preliminary treatment report, generates a corresponding detection number, writes the detection number into the packaged file as a file header, writes the private key into the tail of the packaged file, encrypts the packaged file by using md5, and stores the encrypted file into thestorage module 208 of the cloud platform, so that the data can be extracted through thecloud platform 2 when needed subsequently.
In a preferred embodiment, the image recognition model comprises an image classification model and a plurality of section recognition models, the input end of the image classification model is used as the input end of the image recognition model, the input end of each section recognition model is respectively connected with the output end of the image classification model, and the output ends of all the section recognition models form the output end of the image recognition model;
the image classification model is used for extracting corresponding image characteristics according to an input ultrasonic image, classifying the detection parts of the ultrasonic image according to the image characteristics and outputting a classification result;
each subsection identification model corresponds to a detection part, the image classification model inputs the classification result into the corresponding subsection identification model, and a second class of keywords are obtained through the identification of the subsection identification model and output;
thecloud platform 2 further includes:
and thetraining module 207 is connected with theimage recognition module 202, and thetraining module 207 trains the image recognition model through different types of focus ultrasonic images prestored in thecloud platform 2.
Specifically, the image recognition model is a neural network model established based on a MoileNet neural network, and is trained through different types of focus ultrasonic images prestored in thecloud platform 2, so that a plurality of segmental recognition models are generated in the image recognition model, each segmental recognition model corresponds to one detection part, for example, the current detection part is a thyroid gland, a segmental recognition module corresponding to the thyroid gland is used, and if the current detection part is a liver, respective recognition models corresponding to the liver are used. Further, the lesion features of the current ultrasound image are identified through the corresponding segmentation identification model to extract a second class of keywords. Through the training process of thetraining module 207, the image recognition model can be suitable for various different application scenes, and can recognize ultrasonic images of different detection parts and different lesion characteristics.
The invention also comprises an ultrasonic detection method based on the cloud platform, which is applied to the ultrasonic detection system and comprises the following steps as shown in figure 3:
step S1, the detection personnel inputs the associated information of the patient into the ultrasonic workstation, and the associated information comprises the current detection part;
step S2, the inspector uses the ultrasonic workstation to collect the ultrasonic image of the current inspection part;
step S3, the ultrasound workstation packages the associated information and the ultrasound image into an ultrasound data file, stores the associated information in the file header of the ultrasound data file, and then sends the ultrasound data file to the cloud platform;
step S4, after receiving the ultrasonic data file, the cloud platform carries out semantic segmentation on the associated information in the ultrasonic data file by adopting a semantic recognition model so as to extract a detection part, thereby generating a first class of keywords corresponding to the detection part;
step S5, the cloud platform adopts an image recognition model to extract image features of the ultrasonic image in the ultrasonic data file, so as to generate corresponding second keywords according to the extracted image features;
step S6, fusing the first class keywords and the second class keywords by the cloud platform to form core keywords, and searching corresponding ultrasonic visible entries and ultrasonic suggested entries in an expert word library preset by the cloud platform according to the core keywords;
and step S7, generating at least one preliminary treatment report by the cloud platform ultrasonic seeing terms and the ultrasonic suggestion terms, and pushing the preliminary treatment report to the ultrasonic workstation.
Specifically, in this embodiment, the cloud platform can automatically extract keywords from the associated information and the ultrasound image in the ultrasound data file through a preset neural network model, so as to search corresponding ultrasound term and ultrasound suggested term, and finally form a plurality of preliminary processing reports according to the searched ultrasound term and ultrasound suggested term and feed back to the ultrasound workstation, and a detector can select a report with the best quality from the plurality of preliminary processing reports as a final ultrasound report and feed back the ultrasound report to a patient, and meanwhile, the report can be uploaded to the cloud platform for storage, so as to be extracted when needed subsequently, and by using the method, the time of the detector can be effectively saved, thereby improving the detection efficiency. In addition, a large amount of ultrasonic image data formed by the ultrasonic detection system can be stored in the cloud platform, so that local storage resources are saved.
The technical scheme of the invention has the beneficial effects that: the invention provides an ultrasonic detection system and an ultrasonic detection method based on a cloud platform, wherein the cloud platform respectively adopts a semantic recognition model and an image recognition model to extract first-class keywords and second-class keywords, then the first-class keywords and the second-class keywords are fused to form core keywords, and corresponding ultrasonic visible entries and ultrasonic suggested entries are retrieved through the core keywords, so that a preliminary processing report is automatically generated, the workload of detection personnel is reduced, and the detection efficiency of the detection personnel is improved.
While the invention has been described with reference to a preferred embodiment, it will be understood by those skilled in the art that various changes in form and detail may be made therein without departing from the spirit and scope of the invention.
Claims (9)
1. An ultrasonic detection system based on a cloud platform is characterized by comprising an ultrasonic workstation and a cloud platform, wherein the ultrasonic workstation is in bidirectional communication connection with the cloud platform and comprises:
an entry module for entering associated information of the patient, the associated information including detection position information indicating a current detection site;
the acquisition module is used for acquiring an ultrasonic image of the detection part;
the processing module is respectively connected with the input module and the acquisition module and is used for packaging the associated information and the ultrasonic image into an ultrasonic data file and then sending the ultrasonic data file to the cloud platform;
the cloud platform includes:
a semantic recognition module, which adopts a semantic recognition model formed by pre-training to perform semantic segmentation and recognition on the associated information in the ultrasonic data file, extracts the detection position information and generates a first class of keywords corresponding to the detection part;
the image recognition module is used for extracting image characteristics of the ultrasonic image in the ultrasonic data file by adopting an image recognition model formed by pre-training and generating a corresponding second class keyword according to the extracted image characteristics;
the entry retrieval module is respectively connected with the semantic recognition module and the image recognition module and is used for fusing the first class of keywords and the second class of keywords to form a core keyword and searching corresponding ultrasonic visible entries and ultrasonic suggested entries in an expert word library preset by the cloud platform according to the core keyword;
and the report generating module is connected with the entry retrieval module and used for generating at least one primary processing report according to the ultrasonic visible entry and the ultrasonic suggested entry obtained by searching and feeding the primary processing report back to the ultrasonic workstation.
2. The ultrasonic detection system of claim 1, wherein the processing module comprises:
the file generating unit is used for packaging the associated information and the ultrasonic image to form the ultrasonic data file and storing the associated information in a file header of the ultrasonic data file;
and the format conversion unit is connected with the file generation unit, converts the ultrasonic data file from a first image format to a second image format, and sends the ultrasonic data file converted into the second image format to the cloud platform.
3. The ultrasonic inspection system of claim 1, wherein the ultrasonic workstation further comprises:
and the first encryption module is connected with the processing module and used for encrypting the processed ultrasonic data file and sending the encrypted ultrasonic data file to the cloud platform.
4. The ultrasonic detection system of claim 3, wherein the cloud platform further comprises:
and the decryption module is used for decrypting the ultrasonic data file uploaded by the ultrasonic workstation.
5. The ultrasonic detection system of claim 1, wherein the cloud platform further comprises:
and the second encryption module is used for packaging the ultrasonic data file uploaded by the ultrasonic workstation and the primary processing report generated by the cloud platform, carrying out second encryption on the packaged ultrasonic data file and the primary processing report, and storing the encrypted ultrasonic data file and the encrypted primary processing report in the cloud platform.
6. The ultrasonic detection system according to claim 1, wherein the image recognition model comprises an image classification model and a plurality of section recognition models, an input end of the image classification model is used as an input end of the image recognition model, an input end of each section recognition model is respectively connected with an output end of the image classification model, and output ends of all the section recognition models form an output end of the image recognition model;
the image classification model is used for extracting corresponding image characteristics according to the input ultrasonic image, classifying the detection part of the ultrasonic image according to the image characteristics and outputting a classification result;
each subsection identification model corresponds to one detection part, the image classification model inputs the classification result into the corresponding subsection identification model, and the second class of keywords are obtained through the identification of the subsection identification model and output;
the cloud platform further comprises:
and the training module is connected with the image recognition module and trains the image recognition model through different kinds of focus ultrasonic images prestored in the cloud platform.
7. The ultrasonic testing system of claim 2, wherein the first image format is a JPG format or a BMP format.
8. The ultrasonic inspection system of claim 2, wherein the second image format is a DICOM format.
9. An ultrasonic detection method based on a cloud platform is applied to the ultrasonic detection system of claims 1-8, and comprises the following steps:
step S1, the detection personnel inputs the associated information of the patient into the ultrasonic workstation, and the associated information comprises the current detection part;
step S2, the inspector uses the ultrasonic workstation to acquire the current ultrasonic image of the detected part;
step S3, the ultrasound workstation packages the associated information and the ultrasound image into the ultrasound data file, stores the associated information in a file header of the ultrasound data file, and then sends the ultrasound data file to the cloud platform;
step S4, after receiving the ultrasonic data file, the cloud platform performs semantic segmentation on the associated information in the ultrasonic data file by using the semantic recognition model to extract the detection part, so as to generate the first class keywords corresponding to the detection part;
step S5, the cloud platform adopts the image recognition model to extract image features of the ultrasonic image in the ultrasonic data file, so as to generate the corresponding second type keywords according to the extracted image features;
step S6, the cloud platform fuses the first category of keywords and the second category of keywords to form the core keywords, and searches the corresponding ultrasonic visible entries and the corresponding ultrasonic suggested entries in an expert word library preset by the cloud platform according to the core keywords;
step S7, the cloud platform generates at least one preliminary treatment report by the ultrasonic visible term and the ultrasonic suggested term, and pushes the preliminary treatment report to the ultrasonic workstation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911319654.XACN110946616A (en) | 2019-12-19 | 2019-12-19 | Ultrasonic detection system and ultrasonic detection method based on cloud platform |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911319654.XACN110946616A (en) | 2019-12-19 | 2019-12-19 | Ultrasonic detection system and ultrasonic detection method based on cloud platform |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110946616Atrue CN110946616A (en) | 2020-04-03 |
Family
ID=69982908
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911319654.XAPendingCN110946616A (en) | 2019-12-19 | 2019-12-19 | Ultrasonic detection system and ultrasonic detection method based on cloud platform |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110946616A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113576529A (en)* | 2021-08-31 | 2021-11-02 | 深圳迈瑞动物医疗科技有限公司 | Ultrasonic imaging equipment and code scanning operation method thereof |
| CN114727117A (en)* | 2022-03-04 | 2022-07-08 | 上海深至信息科技有限公司 | Packing method of ultrasonic scanning video |
| CN119671391A (en)* | 2024-12-03 | 2025-03-21 | 北京大学第一医院(北京大学第一临床医学院) | Method for generating basic teaching function evaluation report for medical teachers |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20020077547A1 (en)* | 2000-12-15 | 2002-06-20 | Doug Sluis | Data entry and setup system and method for ultrasound imaging |
| US20050049495A1 (en)* | 2003-09-03 | 2005-03-03 | Siemens Medical Solutions Usa, Inc. | Remote assistance for medical diagnostic ultrasound |
| US20050288584A1 (en)* | 2000-07-21 | 2005-12-29 | Diagnostic Ultrasound Corporation | System for remote evaluation of ultrasound information obtained by a programmed application-specific data collection device |
| US20090310836A1 (en)* | 2008-06-12 | 2009-12-17 | Siemens Medical Solutions Usa, Inc. | Automatic Learning of Image Features to Predict Disease |
| CN102365641A (en)* | 2009-03-26 | 2012-02-29 | 皇家飞利浦电子股份有限公司 | A system that automatically retrieves report templates based on diagnostic information |
| US20120209833A1 (en)* | 2011-02-11 | 2012-08-16 | Siemens Aktiengesellschaft | Methods and devices for data retrieval |
| CN107729319A (en)* | 2017-10-18 | 2018-02-23 | 百度在线网络技术（北京）有限公司 | Method and apparatus for output information |
| US20180107801A1 (en)* | 2016-10-17 | 2018-04-19 | International Business Machines Corporation | Automatic disease detection from unstructured textual reports |
| US20180240551A1 (en)* | 2017-02-20 | 2018-08-23 | General Electric Company | Methods and systems for hierarchical machine learning models for medical imaging |
- 2019
- 2019-12-19CNCN201911319654.XApatent/CN110946616A/enactivePending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050288584A1 (en)* | 2000-07-21 | 2005-12-29 | Diagnostic Ultrasound Corporation | System for remote evaluation of ultrasound information obtained by a programmed application-specific data collection device |
| US20020077547A1 (en)* | 2000-12-15 | 2002-06-20 | Doug Sluis | Data entry and setup system and method for ultrasound imaging |
| US20050049495A1 (en)* | 2003-09-03 | 2005-03-03 | Siemens Medical Solutions Usa, Inc. | Remote assistance for medical diagnostic ultrasound |
| US20090310836A1 (en)* | 2008-06-12 | 2009-12-17 | Siemens Medical Solutions Usa, Inc. | Automatic Learning of Image Features to Predict Disease |
| CN102365641A (en)* | 2009-03-26 | 2012-02-29 | 皇家飞利浦电子股份有限公司 | A system that automatically retrieves report templates based on diagnostic information |
| US20120209833A1 (en)* | 2011-02-11 | 2012-08-16 | Siemens Aktiengesellschaft | Methods and devices for data retrieval |
| US20180107801A1 (en)* | 2016-10-17 | 2018-04-19 | International Business Machines Corporation | Automatic disease detection from unstructured textual reports |
| US20180240551A1 (en)* | 2017-02-20 | 2018-08-23 | General Electric Company | Methods and systems for hierarchical machine learning models for medical imaging |
| CN107729319A (en)* | 2017-10-18 | 2018-02-23 | 百度在线网络技术（北京）有限公司 | Method and apparatus for output information |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113576529A (en)* | 2021-08-31 | 2021-11-02 | 深圳迈瑞动物医疗科技有限公司 | Ultrasonic imaging equipment and code scanning operation method thereof |
| CN113576529B (en)* | 2021-08-31 | 2025-04-01 | 深圳迈瑞动物医疗科技股份有限公司 | Ultrasonic imaging device and code scanning operation method thereof |
| CN114727117A (en)* | 2022-03-04 | 2022-07-08 | 上海深至信息科技有限公司 | Packing method of ultrasonic scanning video |
| CN114727117B (en)* | 2022-03-04 | 2024-07-19 | 上海深至信息科技有限公司 | Packaging method of ultrasonic scanning video |
| CN119671391A (en)* | 2024-12-03 | 2025-03-21 | 北京大学第一医院(北京大学第一临床医学院) | Method for generating basic teaching function evaluation report for medical teachers |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Shen et al. | Privacy-preserving image retrieval for medical IoT systems: A blockchain-based approach | |
| Xia et al. | Towards privacy-preserving content-based image retrieval in cloud computing | |
| KR20230025714A (en) | Document processing and response generation system | |
| JP4169797B2 (en) | Digital authentication center for medical image authentication | |
| CN110946616A (en) | Ultrasonic detection system and ultrasonic detection method based on cloud platform | |
| WO2021108771A1 (en) | Medical imaging processing system | |
| CN110210234B (en) | Medical information migration method and device during referral, computer equipment and storage medium | |
| US9465858B2 (en) | Systems and methods for authenticating and aiding in indexing of and searching for electronic files | |
| CN112202919B (en) | A method and system for storing and retrieving image ciphertext in a cloud storage environment | |
| CN114003929B (en) | Data processing method, device, server and storage medium | |
| CN115705925A (en) | Determining a body region represented by medical imaging data | |
| CN117874144B (en) | Medical data sharing method, device, equipment and storage medium based on blockchain | |
| CN115033917A (en) | Medical data processing system based on privacy protection | |
| CN117711578A (en) | Medical image data analysis management system | |
| Arefeen et al. | irag: Advancing rag for videos with an incremental approach | |
| CN116488890A (en) | Information sharing method and system based on medical community platform | |
| CN115662657A (en) | Online inquiry system based on internet hospital | |
| CN118095442B (en) | Power data tracing method, power system, power equipment and storage medium | |
| CN104737121B (en) | Video playing is multiplexed and is demultiplexed in a browser | |
| CN118824514A (en) | A gastrointestinal early cancer screening and management system based on blockchain technology | |
| CN112447280A (en) | Intelligent medical system for medical image information management | |
| CN117521118A (en) | Medical data searchable encryption privacy protection and system | |
| KR102625319B1 (en) | Method of managing healthcare data based on cloud server and apparatus thereof | |
| KR102354826B1 (en) | Method and system for managing dental clinical picture | |
| CN113051600A (en) | Medical image safety multi-stage desensitization method and system |
Legal Events
| 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 | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right | Effective date of registration:20201015 Address after:Room 5030, 5 / F, building e, 555 Dongchuan Road, Minhang District, Shanghai, 200241 Applicant after:Shanghai Shenzhi Information Technology Co.,Ltd. Address before:Room 3388-c, building 2, 1077 Zuchongzhi Road, China (Shanghai) pilot Free Trade Zone, Pudong New Area, Shanghai Applicant before:Shanghai Zhuxing Biotechnology Co.,Ltd. Applicant before:Shanghai Shenzhi Information Technology Co.,Ltd. | |
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication | Application publication date:20200403 |