A system and method for assisting in identifying and locating pancreatic neuroendocrine tumors

Abstract

The present invention provides a system and method for assisting the identification and localization of pancreatic neuroendocrine tumors, belonging to the field of intelligent assisted tumor diagnosis technology. The tumor-assisted recognition module identifies suspected tumor information in images and displays it in real time. If the recognition match falls below 50%, a guided acquisition instruction is issued to the image identification module. The operator then re-acquires data based on the guided acquisition information, helping less experienced personnel improve detection accuracy and efficiency.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent assisted tumor diagnosis, and specifically provides a system and method for assisting in the identification and positioning of pancreatic neuroendocrine tumors. Background Art

[0002] Pancreatic neuroendocrine neoplasms (pNENs), formerly known as islet cell tumors, have an incidence of approximately 1 to 4 per 100,000 patients, representing approximately 3% of primary pancreatic tumors. Surgery is the mainstay of treatment for pNENs and currently the only potentially curative method. Localization diagnosis plays a crucial role in the surgical treatment of pNENs. Currently, the identification and localization of pancreatic neuroendocrine tumors (PNETs) under endoscopic ultrasound (EUS) still relies on the EUS operator's experience. During EUS scanning, the operator relies on personal experience to identify and locate PNETs within the ultrasound image. However, due to the high similarity between EUS images of PNET lesions and those of the normal pancreas, and the large number of EUS images generated during continuous EUS scanning, the identification and localization of PNETs during the examination is challenging, requiring extensive EUS image interpretation experience and resulting in a certain degree of misdiagnosis and missed diagnosis. Therefore, there is an urgent need for an auxiliary system that can assist the operator in identification and further image acquisition during ultrasound endoscopy scanning. Summary of the Invention

[0003] 1. Technical problem to be solved by the invention

[0004] The purpose of the present invention is to solve the problem that the existing ultrasonic endoscopy detection process requires manual identification and lacks auxiliary equipment.

[0005] 2. Technical solution

[0006] In order to achieve the above object, the technical solution provided by the present invention is:

[0007] A system for assisting in identifying and locating pancreatic neuroendocrine tumors of the present invention comprises:

[0008] An image acquisition module, which is used to obtain image information collected by the endoscope host and send it to the image preprocessing module;

[0009] An image preprocessing module, which is used to perform frame-by-frame screenshot, rotation, cropping, and normalization processing on the image information and then send the processed image data to the tumor auxiliary recognition module;

[0010] A tumor auxiliary recognition module receives processed image data and inputs it into a trained model for recognition, outputs recognition result data to the result display module, or issues a guidance acquisition instruction to the image identification module for data re-acquisition;

[0011] A model training module, comprising a training set unit, a test set unit, a diagnostic model unit, and a correction unit. The training set unit is used to collect and process image data manually identified and annotated by doctors and transmit it to the diagnostic model unit for training. The test set unit is used to transmit the annotated image data to the diagnostic model unit for accuracy testing. The correction unit is used to correct the diagnostic model unit based on the accuracy test results. The tumor auxiliary recognition module operates through the trained diagnostic model unit.

[0012] An image identification module is used to identify the image identified by the model training module, marking it as an analysis result display or marking a re-collected image area;

[0013] A result display module is used to display the identification results of the image identification module and provide real-time editing for doctors.

[0014] Preferably, the image acquisition module acquires image information specifically by identifying it through a fuzzy matching unit, and when a relevant organ is identified, it is automatically acquired after fixed-point marking or manually started by an operator; the frequency of acquiring image information is frame by frame, and the fixed-point markings are specifically a number of organ edge feature points at a certain distance.

[0015] Preferably, the image pre-processing module is rotated specifically by using a fixed point marker as a coincidence point, so that recognition angles of temporally adjacent images are the same.

[0016] Preferably, the diagnostic model unit specifically performs block-by-block hierarchical recognition on the image data, combines pictures from adjacent time periods to obtain block-by-block hierarchical extension directions, and performs focused recognition based on the extension directions.

[0017] Preferably, the tumor auxiliary recognition module sends a guide acquisition instruction to the image identification module for data re-acquisition.

[0018] When the diagnostic model unit obtains the block hierarchical extension direction and fails to obtain accurate data through key identification in the extension direction, the route is marked in the block hierarchical extension direction and sent to the image identification module for visualization. The visualization is specifically to indicate the direction of image recapture through color or arrows.

[0019] Preferably, the image data is divided into blocks for hierarchical identification, specifically by grading the regions according to color differences.

[0020] A method for assisting in identifying and localizing pancreatic neuroendocrine tumors, the method using the above-mentioned system, comprising the following steps:

[0021] S100, collecting image data;

[0022] S200, image data preprocessing;

[0023] S300, performing tumor-assisted recognition on the image data;

[0024] S400: Outputting an auxiliary identification result when relevant information is recognized, or outputting re-collection guidance information when no relevant information is recognized.

[0025] Preferably, the tumor auxiliary recognition module outputs recognition result data mainly including familiarity matching information of various types of tumors and sorting them, and also includes the next step of recognition or diagnosis precautions corresponding to the tumor information.

[0026] Preferably, the familiarity matching information includes matching coincidence rate, size and shape data, matching reason and matching history data, and the matching history data is a number of data with high similarity in previous tumor images of the same type.

[0027] 3. Beneficial effects

[0028] Compared with the prior art, the technical solution provided by the present invention has the following beneficial effects:

[0029] The present invention provides a system for assisting in the identification and localization of pancreatic neuroendocrine tumors. Through a tumor-assisted recognition module, suspected tumor information in an image can be identified and displayed in real time. When the recognition match degree is less than 50%, a guided acquisition instruction is issued to the image identification module. The operator then re-collects data based on the guided acquisition information, helping less experienced personnel improve detection accuracy and efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 This is a flow chart of a method for assisting in identifying and locating pancreatic neuroendocrine tumors according to the present invention. DETAILED DESCRIPTION

[0031] In order to enable those skilled in the art to better understand the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of this application.

[0032] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present application described here. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0033] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in this application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0034] Example 1

[0035] Refer to the attached Figure 1 , a system for assisting in identifying and locating pancreatic neuroendocrine tumors of this embodiment includes:

[0036] An image acquisition module, which is used to obtain image information collected by the endoscope host and send it to the image preprocessing module;

[0037] An image preprocessing module, which is used to perform frame-by-frame screenshot, rotation, cropping, and normalization processing on the image information and then send the processed image data to the tumor auxiliary recognition module;

[0038] A tumor auxiliary recognition module receives processed image data and inputs it into a trained model for recognition, outputs recognition result data to the result display module, or issues a guidance acquisition instruction to the image identification module for data re-acquisition;

[0039] A model training module, comprising a training set unit, a test set unit, a diagnostic model unit, and a correction unit. The training set unit is used to collect and process image data manually identified and annotated by doctors and transmit it to the diagnostic model unit for training. The test set unit is used to transmit the annotated image data to the diagnostic model unit for accuracy testing. The correction unit is used to correct the diagnostic model unit based on the accuracy test results. The tumor auxiliary recognition module operates through the trained diagnostic model unit.

[0040] An image identification module is used to identify the image identified by the model training module, marking it as an analysis result display or marking a re-collected image area;

[0041] A result display module is used to display the identification results of the image identification module and provide real-time editing for doctors.

[0042] The system of the present invention can identify suspected tumor information in images and display it in real time through the tumor auxiliary recognition module. When the recognition matching degree is lower than 50%, a guided acquisition instruction is issued to the image identification module. The operator re-collects data by following the guided acquisition information, helping less experienced personnel improve detection accuracy and efficiency.

[0043] The image acquisition module collects image information specifically by identifying it through a fuzzy matching unit. When the relevant organ is identified, it is automatically collected after fixed-point marking or manually started by the operator; the frequency of collecting image information is frame by frame, and the fixed-point marking is specifically a number of organ edge feature points at a certain distance.

[0044] The rotation of the image pre-processing module is specifically performed by rotating with the fixed point mark as the coincidence point, so that the recognition angles of temporally adjacent images are the same.

[0045] The diagnostic model unit specifically performs block-by-block hierarchical recognition on the image data, combines pictures from adjacent time periods to obtain block-by-block hierarchical extension directions, and performs focused recognition based on the extension directions.

[0046] The tumor auxiliary recognition module sends a guide acquisition instruction to the image identification module to re-acquire data.

[0047] When the diagnostic model unit obtains the block hierarchical extension direction and fails to obtain accurate data through key identification in the extension direction, the route is marked in the block hierarchical extension direction and sent to the image identification module for visualization. The visualization is specifically to indicate the direction of image recapture through color or arrows.

[0048] The image data is divided into blocks for hierarchical recognition, specifically for regional classification based on color difference.

[0049] Example 2

[0050] Refer to the attached Figure 1 This embodiment provides a method for assisting in identifying and locating pancreatic neuroendocrine tumors. The method comprises the system described in Example 1, and the method includes the following steps:

[0051] S100, collecting image data;

[0052] S200, image data preprocessing;

[0053] S300, performing tumor-assisted recognition on the image data;

[0054] S400: Outputting an auxiliary identification result when relevant information is recognized, or outputting re-collection guidance information when no relevant information is recognized.

[0055] The tumor auxiliary recognition module outputs recognition result data mainly including familiarity matching information of various types of tumors and sorting, and also includes the next step of recognition or diagnosis precautions corresponding to the tumor information.

[0056] The familiarity matching information includes matching coincidence rate, size and shape data, matching reason and matching history data, and the matching history data is a number of data with high similarity in previous tumor images of the same type.

[0057] The above-mentioned embodiments only express a certain implementation method of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the patent scope of the present invention. It should be pointed out that for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all fall within the scope of protection of the present invention. Therefore, the scope of protection of the patent of the present invention shall be based on the attached claims.

Claims

1. A system for assisting in the identification and localization of pancreatic neuroendocrine tumors, characterized by: include An image acquisition module, which is used to obtain image information collected by the endoscope host and send it to the image preprocessing module; An image preprocessing module, which is used to perform frame-by-frame screenshot, rotation, cropping, and normalization processing on the image information and then send the processed image data to the tumor auxiliary recognition module; A tumor auxiliary recognition module receives processed image data and inputs it into a trained model for recognition, outputs recognition result data to the result display module, or issues a guidance acquisition instruction to the image identification module for data re-acquisition; A model training module, comprising a training set unit, a test set unit, a diagnostic model unit, and a correction unit. The training set unit is used to collect and process image data manually identified and annotated by doctors and transmit it to the diagnostic model unit for training. The test set unit is used to transmit the annotated image data to the diagnostic model unit for accuracy testing. The correction unit is used to correct the diagnostic model unit based on the accuracy test results. The tumor auxiliary recognition module operates through the trained diagnostic model unit. An image identification module is used to identify the image identified by the model training module, marking it as an analysis result display or marking a re-collected image area; A result display module, which is used to display the identification results of the image identification module and provide real-time editing for doctors; The diagnostic model unit specifically performs block-by-block hierarchical recognition on the image data, combines the images of adjacent time periods to obtain the block-by-block hierarchical extension direction, and performs key recognition based on the extension direction; The tumor auxiliary recognition module sends a guide acquisition instruction to the image identification module to re-acquire data. When the diagnostic model unit obtains the block hierarchical extension direction and fails to obtain accurate data through key identification in the extension direction, the route is marked in the block hierarchical extension direction and sent to the image identification module for visualization. The visualization is specifically to indicate the direction of image re-collection through color or arrows; the image data is specifically to be classified into regions according to color difference.

2. The system for assisting in identifying and localizing pancreatic neuroendocrine tumors according to claim 1, characterized in that: The image acquisition module collects image information specifically by identifying it through a fuzzy matching unit. When the relevant organ is identified, it is automatically collected after fixed-point marking or manually started by the operator; the frequency of collecting image information is frame by frame, and the fixed-point marking is specifically a number of organ edge feature points at a certain distance.

3. The system for assisting in identifying and localizing pancreatic neuroendocrine tumors according to claim 1, characterized in that: The rotation of the image pre-processing module is specifically performed by rotating with the fixed point mark as the coincidence point, so that the recognition angles of temporally adjacent images are the same.

4. A method for assisting in the identification and localization of pancreatic neuroendocrine tumors, characterized in that: The method adopts any of the above-mentioned systems, and the method comprises the following steps: S100, collecting image data; S200, image data preprocessing; S300, performing tumor-assisted recognition on the image data; S400: Outputting an auxiliary identification result when relevant information is recognized, or outputting re-collection guidance information when no relevant information is recognized.

5. The method for assisting in identifying and locating pancreatic neuroendocrine tumors according to claim 4, characterized in that: The tumor auxiliary recognition module outputs recognition result data mainly including familiarity matching information of various types of tumors and sorting, and also includes the next step of recognition or diagnosis precautions corresponding to the tumor information.

6. The method for assisting in identifying and locating pancreatic neuroendocrine tumors according to claim 5, characterized in that: The familiarity matching information includes matching coincidence rate, size and shape data, matching reason and matching history data, and the matching history data is a number of data with high similarity in previous tumor images of the same type.

Images