WO2023087183A1

WO2023087183A1 - Method for typing chamber angle closure mechanism of primary angle closure glaucoma

Info

Publication number: WO2023087183A1
Application number: PCT/CN2021/131258
Authority: WO
Inventors: 张烨; 李树宁; 张青; 张潇月
Original assignee: 首都医科大学附属北京同仁医院
Priority date: 2021-11-17
Filing date: 2021-11-17
Publication date: 2023-05-25

Abstract

The present application discloses a method for typing a chamber angle closure mechanism of primary angle closure glaucoma. The method comprises: obtaining an anterior segment cross-sectional image to be typed; performing chamber angle feature detection on the anterior segment cross-sectional image to obtain a chamber angle feature region image; inputting the chamber angle feature region image into a preset first typing network model to obtain a first typing result; transforming the chamber angle feature region image into a heat map, and inputting the heat map obtained by transformation and the anterior segment cross-sectional image into a preset second typing network model to obtain a second typing result; and analyzing the first typing result and the second typing result to obtain a typing result of the anterior segment cross-sectional image. In the present application, by processing the anterior segment cross-sectional image in two different modes and performing analysis to obtain a final result, the accuracy of typing is improved, and convenience is brought to a doctor for subsequent diagnosis; moreover, typing processing can be quickly performed on the image by means of the typing network models, thereby improving the efficiency of typing.

Description

A classification method of angle closure mechanism in primary angle-closure glaucoma

technical field

The present application relates to the technical field of medical image processing, in particular to a method for typing the angle closure mechanism of primary angle-closure glaucoma.

Background technique

Glaucoma is the first irreversible blinding eye disease worldwide. Among them, the number of patients with primary angle-closure glaucoma (abbreviated as closed glaucoma) ranks first in the world in my country. The pathogenesis of closed eyes in my country is different from that of Westerners. In addition to the pupillary block mechanism, there are still non-pupil block factors in the angle closure mechanism, which is diverse, and in some cases, two mechanisms or more mechanisms may co-exist. The establishment of the classification of angle closure mechanism of closed eyes is helpful for targeted and individualized treatment of closed eyes with different angle closing mechanisms, and it is helpful for the exploration of risk factors for different angle closing mechanisms of closed eyes , and establish a prediction model for the onset and progression of the disease.

The existing classification methods of angle closure mechanism are completed by doctors or researchers, and are highly subjective. In addition, due to the complexity of the chamber angle closure mechanism of patients with closed eyes in my country, in the case of classification mainly through the doctor's subjective judgment, the doctor's judgment process may consume a lot of time, resulting in low efficiency of the entire classification process, which cannot be used for classification. The doctor's diagnosis provides powerful help, and it is not suitable for widespread application in clinical and research.

Contents of the invention

In view of this, the embodiment of the present application provides a method for typing the angle closure mechanism of primary angle-closure glaucoma, so as to solve the problem of low typing efficiency of the existing typing methods.

In order to achieve the above object, the application provides the following technical solutions:

The embodiment of the present application provides a method for typing the angle closure mechanism of primary angle-closure glaucoma, including:

Obtain cross-sectional images of the anterior segment to be classified;

Performing a feature detection of the room angle on the anterior segment cross-sectional image to obtain a room angle feature region image;

Inputting the image of the characteristic area of the chamber angle into the preset first classification network model to obtain the first classification result;

Transforming the image of the characteristic area of the chamber angle into a thermal map, and inputting the converted thermal image and the anterior segment cross-sectional image into a preset second classification network model to obtain a second classification result;

The first typing result and the second typing result are analyzed to obtain the typing result of the anterior segment cross-sectional image.

Optionally, the feature detection of the room angle is performed on the cross-sectional image of the anterior segment to obtain the image of the room angle feature area, including:

Performing a feature detection of the room angle on the cross-sectional image of the anterior segment to obtain a first room angle feature area image and a second room angle feature area image;

Combining the first image of the characteristic region of the chamber angle with the image of the second characteristic region of the chamber angle to obtain the image of the characteristic region of the chamber angle.

Optionally, the feature detection of the room angle is performed on the cross-sectional image of the anterior segment to obtain the first room angle feature area image and the second room angle feature area image, including:

Performing feature detection of the chamber angle on the cross-sectional image of the anterior segment to obtain a first chamber angle region and a second chamber angle region;

The anterior segment cross-sectional image is clipped based on the first chamber angle region and the second chamber angle region to obtain a first chamber angle characteristic region image and a second chamber angle characteristic region image.

Optionally, before clipping the anterior segment cross-sectional image based on the first chamber angle region and the second chamber angle region, the method further includes:

Acquiring sizes corresponding to the first chamber angle characteristic area and the second chamber angle characteristic area;

Comparing the sizes of the first chamber angle characteristic region and the second chamber angle characteristic region;

According to the comparison result, the small-sized chamber angle characteristic region is adjusted so that the sizes of the first chamber angle characteristic region and the second chamber angle characteristic region are consistent.

Optionally, converting the image of the characteristic area of the chamber angle into a thermal map, and inputting the converted thermal image and the cross-sectional image of the anterior segment into a preset second classification network model to obtain a second classification result, include:

performing Gaussian transformation on the image of the chamber angle feature region to generate a Gaussian distribution heat map;

Input the Gaussian distribution heat map and the anterior segment cross-sectional image into the preset second classification network model to obtain the second classification result.

Optionally, analyzing the first typing result and the second typing result to obtain the typing result of the anterior segment cross-sectional image, including:

Performing a weighted comparative analysis on the first typing result and the second typing result to obtain the probability values corresponding to each typing category;

The typing category with the highest probability value is determined as the typing result of the anterior segment cross-sectional image.

The embodiment of the present application also provides an angle-closure mechanism classification device for primary angle-closure glaucoma, including:

Acquisition module: used to acquire cross-sectional images of the anterior segment to be classified;

Detection module: used to perform feature detection of the room angle on the anterior segment cross-sectional image, and obtain an image of the room angle feature region;

The first classification module: for inputting the image of the characteristic area of the chamber angle into the preset first classification network model to obtain the first classification result;

The second classification module: used to convert the thermal map of the image of the characteristic region of the room angle, and input the converted thermal image and the cross-sectional image of the anterior segment into the preset second classification network model to obtain the second Typing results;

An analysis module: used for performing a feature detection of the room angle on the cross-sectional image of the anterior segment to obtain an image of a room angle feature area.

Optionally, the second typing module includes:

Transformation module: used to perform Gaussian transformation on the image of the room angle feature region to generate a Gaussian distribution heat map;

Typing processing module: used to input the Gaussian distribution heat map and the anterior segment cross-sectional image into the preset second typing network model to obtain the second typing result.

The embodiment of the present application also provides an electronic device, including:

A memory and a processor, the memory and the processor are connected in communication with each other, computer instructions are stored in the memory, and the processor executes the computer instructions to execute the originality provided by the embodiments of the present application. Classification of angle-closure mechanism in angle-closure glaucoma.

The embodiment of the present application also provides a computer-readable storage medium, the computer-readable storage medium stores computer instructions, and the computer instructions are used to make the computer execute the primary angle-closure glaucoma method provided in the embodiment of the present application. A classification method for the angle closure mechanism.

The technical solution of the present application has the following advantages:

This application provides a method for typing the angle closure mechanism of primary angle-closure glaucoma, by obtaining the cross-sectional image of the anterior segment to be typed; performing a characteristic detection of the angle on the cross-sectional image of the anterior segment to obtain the angle Feature area image; input the image of the room angle feature area into the preset first classification network model to obtain the first classification result; convert the image of the room angle feature area into a thermal map, and transect the transformed thermal image with the anterior segment The surface image is input into the preset second classification network model to obtain the second classification result; the first classification result and the second classification result are analyzed to obtain the classification result of the anterior segment cross-sectional image. This application realizes the rapid and objective classification of closed angle closure mechanism based on artificial intelligence. The cross-sectional images of the anterior segment are processed in two different ways to obtain two classification results, and then through the two classification results Perform analysis to get the final result. Compared with the traditional manual typing method, the two typing results are more objective and accurate, which effectively improves the typing accuracy and brings convenience to the doctor's subsequent diagnosis. At the same time, through the typing network The model can quickly classify images and improve the efficiency of classification.

Description of drawings

In order to more clearly illustrate the specific embodiments of the present application or the technical solutions in the prior art, the following will briefly introduce the accompanying drawings that need to be used in the description of the specific embodiments or prior art. Obviously, the accompanying drawings in the following description The figures show some implementations of the present application, and those skilled in the art can obtain other figures based on these figures without any creative effort.

Fig. 1 is the flow chart of the angle closure mechanism classification method of primary angle-closure glaucoma in the embodiment of the present application;

Fig. 2 is a flow chart of obtaining an image of a room angle feature region according to an embodiment of the present application;

Fig. 3 is a flow chart of feature detection of room angle on an anterior segment cross-sectional image according to an embodiment of the present application;

Fig. 4 is a flow chart of adjusting the image of the room angle feature region according to the embodiment of the present application;

Fig. 5 is a flow chart of obtaining the second typing result according to the embodiment of the present application;

FIG. 6 is a flow chart of weighted comparative analysis of typing results according to the embodiment of the present application;

FIG. 7 is a schematic structural diagram of an angle-closure mechanism typing device for primary angle-closure glaucoma in an embodiment of the present application;

Fig. 8 is a schematic structural diagram of the second typing module in the embodiment of the present application;

FIG. 9 is a schematic structural diagram of an electronic device in an embodiment of the present application.

Detailed ways

In order to make the purposes, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments It is a part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without making creative efforts belong to the scope of protection of this application.

According to the embodiment of the present application, an embodiment of a method for typing the angle closure mechanism of primary angle-closure glaucoma is provided. Instructions are executed in a computer system and, although a logical order is shown in the flowcharts, in some cases the steps shown or described may be performed in an order different from that shown or described herein.

In this embodiment, a method for typing the angle closure mechanism of primary angle-closure glaucoma is provided, as shown in Figure 1, the method for typing the angle closure mechanism of primary angle-closure glaucoma includes the following steps :

Step S1: Obtain cross-sectional images of the anterior segment to be classified. Specifically, the cross-sectional image of the anterior segment to be classified is an optical coherence tomography (OCT) image of the cross-section of the anterior segment. Anterior segment OCT examination using invisible light can obtain a clear cross-sectional image of the anterior segment, and can Non-contact, qualitative/quantitative assessment and measurement of the anterior chamber and angle. At the same time, the anterior segment OCT avoids the impact on the angle shape and pupil diameter, and does not compress the eyeball, so that the existing angle closure will not be missed due to the interference of the inspection operation.

Step S2: The feature detection of the room angle is performed on the cross-sectional image of the anterior segment to obtain the image of the room angle feature area. Specifically, the Roi detection network is used to detect the feature of the anterior segment cross-sectional image of the room angle, obtain the image near the room angle that is focused on and frame it, and obtain the image of the room angle feature area. The room angle feature area obtained in this way It is more accurate and provides guarantee for the subsequent typing process.

Step S3: Input the image of the characteristic area of the chamber angle into the preset first classification network model to obtain the first classification result. Specifically, the first classification network model is a convolutional neural network. The image of the room angle feature region in this process is a feature map with global information and enhanced local information. By inputting the image of the room angle feature region into the preset first The one-category network model can obtain a more objective first-category result than traditional manual judgment.

Step S4: Transform the image of the characteristic area of the chamber angle into a thermal map, and input the converted thermal image and the cross-sectional image of the anterior segment into the preset second classification network model to obtain the second classification result. Specifically, the deep learning network model of the second classification network model is obtained after artificial intelligence training through a large number of case images and diagnosis results. The Gaussian distribution heat map is obtained by converting the heat force, and the Gaussian distribution heat map and the anterior segment are transected. Input the deep learning network model together with the surface image, so that the deep learning network model can analyze the global information of the picture and focus on the Gaussian circle area of the thermal image to obtain the second classification result.

Step S5: Analyzing the first typing result and the second typing result to obtain the typing result of the cross-sectional image of the anterior segment.

Through the above steps S1 to S5, the method for typing the angle closure mechanism of primary angle-closure glaucoma provided by the embodiment of the present application, the application realizes the rapid and objective classification of the angle closure mechanism based on artificial intelligence , the cross-sectional image of the anterior segment is processed in two different ways to obtain two typing results, and then the final result is obtained by analyzing the two typing results. Compared with the traditional manual typing method, the two typing results , are more objective and accurate, effectively improving the accuracy of typing, and bringing convenience to doctors' follow-up diagnosis; at the same time, through the typing network model, images can be typed quickly, improving the efficiency of typing.

Specifically, in one embodiment, the above step S2, as shown in FIG. 2 , specifically includes the following steps:

Step S21: Perform feature detection of the room angle on the cross-sectional image of the anterior segment to obtain a first room angle feature area image and a second room angle feature area image. Specifically, in the cross-sectional image of the anterior segment, the characteristic regions of the chamber angle are divided into left and right sides, and the characteristic regions need to be extracted separately.

Step S22: combining the first image of the characteristic region of the chamber angle and the image of the second characteristic region of the chamber angle to obtain the image of the characteristic region of the chamber angle.

Specifically, in one embodiment, the above step S21, as shown in FIG. 3 , specifically includes the following steps:

Step S211: Perform feature detection of the chamber angle on the cross-sectional image of the anterior segment to obtain the first chamber angle region and the second chamber angle region. Specifically, before extracting the room angle characteristic area, the position of the room angle area needs to be determined. Through the target detection of the ROI, the position of the area to be placed can be obtained more accurately, and the feature area is framed.

Step S212: Clipping the anterior segment cross-sectional image based on the first chamber angle region and the second chamber angle region to obtain the first chamber angle characteristic region image and the second chamber angle characteristic region image. Specifically, what is obtained after clipping is an image of a room angle feature region with global information and enhanced local information.

Specifically, in one embodiment, before performing the above-mentioned step S212, as shown in FIG. 4 , the following steps are specifically included:

Step S2101: Acquiring the corresponding sizes of the first chamber angle characteristic area and the second chamber angle characteristic area;

Step S2102: Comparing the sizes of the first chamber angle characteristic area and the second chamber angle characteristic area;

Step S2103: According to the comparison result, adjust the small-sized chamber angle characteristic region, so that the sizes of the first chamber angle characteristic region and the second chamber angle characteristic region are the same.

Specifically, during the process of extracting the image of the characteristic region of the chamber angle, the sizes of the first characteristic region image of the chamber angle and the image of the second characteristic region of the chamber angle on the left and right sides may be inconsistent in size, and need to be adjusted. The adjustment process can be adjusted through ROI pooling, or adjusted in other ways to make the first chamber angle characteristic region image and the second chamber angle characteristic region image have the same size, and no specific limitation is set here. The image of the characteristic area of the chamber angle with consistent size facilitates the subsequent classification.

Specifically, in one embodiment, the above step S4, as shown in FIG. 5 , specifically includes the following steps:

Step S41: Gaussian transformation is performed on the image of the characteristic region of the room angle to generate a Gaussian distribution heat map. Specifically, the Gaussian distribution heat map reflects the probability distribution of the key attention parts.

Step S42: Input the Gaussian distribution heat map and the cross-sectional image of the anterior segment into the preset second classification network model to obtain the second classification result. Specifically, the deep learning network model of the second classification network model can be replaced by other deep learning classification networks, including but not limited to: inceptionV4, seResnet, etc. It is obtained after artificial intelligence training through a large number of case images and diagnostic results. By inputting the Gaussian distribution heat map and the anterior segment cross-sectional image together into the deep learning network model, the deep learning network model can analyze the global information of the picture and at the same time focus on Focus on the Gaussian circle area of the thermal image to get the second classification result, which makes the classification result more objective and accurate.

Specifically, in one embodiment, the above step S5, as shown in FIG. 6, specifically includes the following steps:

Step S51: Perform weighted comparative analysis on the first typing result and the second typing result to obtain the probability values corresponding to each typing category. Specifically, when performing weighted comparative analysis, logit weighting, softmax weighting, 01 weighting, etc. can be selected.

Step S52: Determine the typing category with the highest probability value as the typing result of the anterior segment cross-sectional image. Specifically, taking softmax weighting as an example, using the probability values of the first typing result and the second typing result as the judgment basis, a richer amount of information can be analyzed to obtain a more balanced typing result.

In this application, the cross-sectional images of the anterior segment are processed in two different ways to obtain two typing results, and then the final results are obtained by analyzing the two typing results. Compared with the traditional manual typing, the two typing results The method is more objective and accurate, which effectively improves the accuracy of typing, effectively improves the accuracy of typing, and brings convenience to doctors' follow-up diagnosis. At the same time, images can be typed quickly through the typing network model. , improving the typing efficiency.

This embodiment also provides an angle-closure mechanism classification device for primary angle-closure glaucoma, which is used to implement the above-mentioned embodiments and preferred implementation modes, and those that have already been described will not be repeated. As used below, the term "module" may be a combination of software and/or hardware that realizes a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, implementations in hardware, or a combination of software and hardware are also possible and contemplated.

The present embodiment provides an angle-closure mechanism typing device for primary angle-closure glaucoma, as shown in FIG. 7 , comprising:

The acquiring module 101 is configured to acquire the cross-sectional image of the anterior segment to be typed. For details, refer to the relevant description of step S1 in the above method embodiment, which will not be repeated here.

The detection module 102 is configured to detect the feature of the room angle on the cross-sectional image of the anterior segment to obtain the image of the feature area of the room angle. For details, refer to the relevant description of step S2 in the above method embodiment, and will not be repeated here.

The first classification module 103 is configured to input the image of the characteristic area of the chamber angle into the preset first classification network model to obtain the first classification result. For details, refer to the relevant description of step S3 in the above method embodiment, here No further details will be given.

The second typing module 104 is used to convert the thermal map of the image of the characteristic area of the room angle, and input the converted thermal image and the cross-sectional image of the anterior segment into the preset second typing network model to obtain the second classification network model. For the dichotomous results, refer to the relevant description of step S4 in the above method embodiment for details, and details are not repeated here.

The analysis module 105 performs feature detection of the room angle on the anterior segment cross-sectional image to obtain an image of the room angle characteristic area. For details, refer to the relevant description of step S5 in the above method embodiment, and will not be repeated here.

Specifically, in one embodiment, the above-mentioned second typing module 104, as shown in FIG. 8 , specifically includes:

Transformation module 1041: used to perform Gaussian transformation on the image of the characteristic region of the room angle to generate a heat map of Gaussian distribution. For details, refer to the relevant description of step S41 in the above method embodiment, which will not be repeated here.

Typing processing module 1042: used to input the Gaussian distribution heat map and the anterior segment cross-sectional image into the preset second typing network model to obtain the second typing result. For details, refer to the relevant description of step S42 in the above method embodiment , which will not be repeated here.

The angle-closure mechanism classification device for primary angle-closure glaucoma in this embodiment is presented in the form of a functional unit, where the unit refers to an ASIC circuit, a processor that executes one or more software or fixed programs, and Memory, and/or other devices that can provide the above functions.

Further functional descriptions of the above-mentioned modules are the same as those in the above-mentioned corresponding embodiments, and will not be repeated here.

According to the embodiment of the present application, an electronic device is also provided. As shown in FIG. 9, the electronic device may include a processor 901 and a memory 902, wherein the processor 901 and the memory 902 may be connected through a bus or in other ways. In FIG. 9, the Take connection via bus as an example.

The processor 901 may be a central processing unit (Central Processing Unit, CPU). The processor 901 can also be other general-purpose processors, digital signal processors (Digital Signal Processor, DSP), application-specific integrated circuits (Application Specific Integrated Circuit, ASIC), field-programmable gate array (Field-Programmable Gate Array, FPGA) or Other chips such as programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or combinations of the above-mentioned types of chips.

As a non-transitory computer-readable storage medium, the memory 902 can be used to store non-transitory software programs, non-transitory computer-executable programs and modules, such as program instructions/modules corresponding to the methods in the method embodiments of the present application. The processor 901 executes various functional applications and data processing of the processor by running the non-transitory software programs, instructions and modules stored in the memory 902, that is, implements the methods in the above method embodiments.

The memory 902 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and an application program required by at least one function; the data storage area may store data created by the processor 901 and the like. In addition, the memory 902 may include a high-speed random access memory, and may also include a non-transitory memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state storage devices. In some embodiments, the storage 902 may optionally include storages that are remotely located relative to the processor 901, and these remote storages may be connected to the processor 901 through a network. Examples of the aforementioned networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

One or more modules are stored in the memory 902, and when executed by the processor 901, the methods in the foregoing method embodiments are executed.

The specific details of the above-mentioned electronic device can be understood by correspondingly referring to the corresponding descriptions and effects in the above-mentioned method embodiments, and will not be repeated here.

Those skilled in the art can understand that all or part of the processes in the methods of the above-mentioned embodiments can be completed by instructing related hardware through computer programs, and the programs can be stored in a computer-readable storage medium. , may include the flow of the embodiments of the above-mentioned methods. Wherein, the storage medium can be a magnetic disk, an optical disk, a read-only memory (Read-Only Memory, ROM), a random access memory (Random Access Memory, RAM), a flash memory (Flash Memory), a hard disk (Hard Disk Drive) , abbreviation: HDD) or solid-state drive (Solid-State Drive, SSD), etc.; the storage medium may also include a combination of the above-mentioned types of memory.

Although the embodiment of the application has been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the application, and such modifications and variations all fall within the scope of the appended claims. within the bounds of the requirements.

Claims

A method for typing the angle closure mechanism of primary angle-closure glaucoma, comprising:

Obtain cross-sectional images of the anterior segment to be classified;

Performing a feature detection of the room angle on the anterior segment cross-sectional image to obtain a room angle feature region image;

Inputting the image of the characteristic area of the chamber angle into the preset first classification network model to obtain the first classification result;

Transforming the image of the characteristic area of the chamber angle into a thermal map, and inputting the converted thermal image and the anterior segment cross-sectional image into a preset second classification network model to obtain a second classification result;

The first typing result and the second typing result are analyzed to obtain the typing result of the anterior segment cross-sectional image.
The method for typing the angle closure mechanism of primary angle-closure glaucoma according to claim 1, wherein the feature detection of the angle is performed on the cross-sectional image of the anterior segment to obtain an image of the characteristic area of the angle ,include:

Performing a feature detection of the room angle on the cross-sectional image of the anterior segment to obtain a first room angle feature area image and a second room angle feature area image;

Combining the first image of the characteristic region of the chamber angle with the image of the second characteristic region of the chamber angle to obtain the image of the characteristic region of the chamber angle.
The method for typing the angle closure mechanism of primary angle-closure glaucoma according to claim 2, wherein the feature detection of the room angle is performed on the anterior segment cross-sectional image to obtain the first room angle feature Regional images and second angle characteristic regional images, including:

Performing feature detection of the chamber angle on the cross-sectional image of the anterior segment to obtain a first chamber angle region and a second chamber angle region;

The anterior segment cross-sectional image is clipped based on the first chamber angle region and the second chamber angle region to obtain a first chamber angle characteristic region image and a second chamber angle characteristic region image.
The angle-closure mechanism classification method for primary angle-closure glaucoma according to claim 3, wherein the anterior segment cross-sectional image is analyzed based on the first angle area and the second angle area Before clipping, also include:

Acquiring sizes corresponding to the first chamber angle characteristic area and the second chamber angle characteristic area;

Comparing the sizes of the first chamber angle characteristic region and the second chamber angle characteristic region;

According to the comparison result, the small-sized chamber angle characteristic region is adjusted so that the sizes of the first chamber angle characteristic region and the second chamber angle characteristic region are consistent.
The method for typing the angle closure mechanism of primary angle-closure glaucoma according to claim 1, wherein the conversion of the thermal map of the image of the characteristic region of the room angle is performed, and the converted thermal image Input the preset second classification network model with the anterior segment cross-sectional image to obtain a second classification result, including:

performing Gaussian transformation on the image of the chamber angle feature region to generate a Gaussian distribution heat map;

Input the Gaussian distribution heat map and the anterior segment cross-sectional image into the preset second classification network model to obtain the second classification result.
The angle-closure mechanism typing method for primary angle-closure glaucoma according to claim 1, wherein the first typing result and the second typing result are analyzed to obtain the The classification results of section cross-sectional images, including:

Performing a weighted comparative analysis on the first typing result and the second typing result to obtain the probability values corresponding to each typing category;

The typing category with the highest probability value is determined as the typing result of the anterior segment cross-sectional image.
An angle-closure mechanism classification device for primary angle-closure glaucoma, characterized in that it includes:

Acquisition module: used to acquire cross-sectional images of the anterior segment to be classified;

Detection module: used to perform feature detection of the room angle on the anterior segment cross-sectional image, and obtain an image of the room angle feature region;

The first classification module: for inputting the image of the characteristic area of the chamber angle into the preset first classification network model to obtain the first classification result;

The second classification module: used to convert the thermal map of the image of the characteristic region of the room angle, and input the converted thermal image and the cross-sectional image of the anterior segment into the preset second classification network model to obtain the second Typing result;

Analysis module: used to analyze the first typing result and the second typing result to obtain the typing result of the anterior segment cross-sectional image.
The angle-closure mechanism typing device for primary angle-closure glaucoma according to claim 7, wherein the second typing module includes:

Transformation module: used to perform Gaussian transformation on the image of the room angle feature region to generate a Gaussian distribution heat map;

Typing processing module: used to input the Gaussian distribution heat map and the anterior segment cross-sectional image into the preset second typing network model to obtain the second typing result.
An electronic device, characterized in that it comprises:

A memory and a processor, the memory and the processor are connected in communication with each other, computer instructions are stored in the memory, and the processor performs any one of claims 1 to 7 by executing the computer instructions The method for typing the angle closure mechanism of primary angle-closure glaucoma.
A computer-readable storage medium, characterized in that the computer-readable storage medium stores computer instructions, and the computer instructions are used to make the computer perform the original invention described in any one of claims 1 to 7. Classification of angle-closure mechanism in angle-closure glaucoma.