Artificial intelligence-based A-type gastritis information processing method and device
By employing an artificial intelligence-based approach, combining convolutional neural networks and large language models with machine learning techniques, we have achieved highly accurate classification of images of type A gastritis, solving the diagnostic difficulties in existing technologies and improving the reliability of diagnosis.
Patent Information
- Application Number
- CN202511118276.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-08-11
- Publication Date
- 2025-11-14
AI Technical Summary
Current technologies for classifying type A gastritis images have low accuracy and a high rate of missed diagnoses, and the lack of experienced physicians makes diagnosis difficult.
An artificial intelligence-based approach was adopted, using a convolutional neural network model to identify the location in gastroscopy images and combining it with a large language model for abnormal feature classification. By comprehensively analyzing image features and baseline feature information through a machine learning model, the accuracy of image classification was improved.
It improved the accuracy of image classification for type A gastritis, reduced the false negative rate, and enhanced the reliability of diagnosis.
Smart Images
Figure CN120953240A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of artificial intelligence technology, specifically to an information processing method and apparatus for type A gastritis based on artificial intelligence. Background Technology
[0002] Type A gastritis, also known as autoimmune gastritis, is a chronic gastritis caused by an autoimmune response. It primarily attacks gastric parietal cells, leading to decreased gastric acid secretion and intrinsic factor deficiency. Studies show that patients with type A gastritis have a 3-5 times higher risk of developing gastric cancer than the general population. Its harm extends beyond the digestive system, potentially causing systemic complications such as pernicious anemia, vitamin B12 deficiency, and related neurological abnormalities. In the middle and late stages, type A gastritis exhibits more severe atrophy compared to type B gastritis, and the risk of developing gastric cancer is significantly higher. Clinically, due to insufficient understanding of type A gastritis and a lack of experienced physicians, the rate of missed diagnoses of type A gastritis is high, and the accuracy of image classification is low. Summary of the Invention
[0003] This application provides an information processing method and apparatus for type A gastritis based on artificial intelligence, which can improve the accuracy of image classification.
[0004] Firstly, the information processing method for type A gastritis based on artificial intelligence provided in this application includes: Acquire multiple gastroscopy images obtained during gastroscopy of the target entity; Based on a preset convolutional neural network model, the location of each gastroscopy image is identified to obtain the location category of the gastroscopy image; For each of the K location categories belonging to the preset location category set, N gastroscopic images belonging to the location category are obtained respectively, resulting in K*N gastroscopic images, where K and N are both positive integers; Based on a pre-defined large language model, each of the K*N gastroscopy images is classified into P abnormal feature categories, and P abnormal classification confidence scores of the gastroscopy image belonging to the P abnormal feature categories are obtained. P abnormal classification confidence scores of each of the K*N gastroscopy images are obtained, where P is a positive integer greater than 1. The gastroscopy image classification category is determined based on P abnormal classification confidence scores for each of the K*N gastroscopy images, wherein the gastroscopy image classification category is a normal category, a first abnormal image category, or a second abnormal image category.
[0005] In an optional embodiment, determining the gastroscopy image classification category based on P abnormality classification confidence scores for each of the K*N gastroscopy images includes: When N is greater than 1, each of the above-mentioned site categories is determined as the first site category, and each of the abnormal feature categories is determined as the first abnormal feature category, so as to obtain N abnormal classification confidence scores of N gastroscopy images under the first site category belonging to the first abnormal feature category; The average of the N abnormal classification confidence scores is determined as the feature confidence score of the gastroscopy image belonging to the first abnormal feature category under the first site category, thus obtaining K*P feature confidence scores of the gastroscopy image belonging to P abnormal feature categories under K site categories; The classification category of gastroscopy images is determined based on the confidence scores of K*P features.
[0006] In an optional embodiment, determining the gastroscopy image classification category based on K*P feature confidence scores includes: Obtain the baseline feature information of the target entity, which includes age and gender features; The baseline feature information of the target entity and the confidence scores of K*P features are input into a preset machine learning model to obtain the classification category of the gastroscopy image.
[0007] In an optional embodiment, the preset machine learning model is a random forest model, a decision tree model, or a support vector machine.
[0008] In an optional embodiment, K is 8, and the K site categories are respectively the gastric antrum category, duodenal bulb category, descending duodenal part category, lower part of the gastric body category (positive endoscope), upper and middle part of the gastric body category (positive endoscope), fundus category (reverse endoscope), upper and middle part of the gastric body category (reverse endoscope), and angle category (reverse endoscope).
[0009] In an optional embodiment, P is 8, and the P abnormal feature categories are old hemorrhage, redness, atrophy, fold swelling, goosebumps, mucosal swelling, intestinal metaplasia, and bile reflux.
[0010] Secondly, the artificial intelligence-based information processing device for type A gastritis provided in this application includes: The first acquisition module is used to acquire multiple gastroscopy images obtained when performing gastroscopy on the target entity; The recognition module is used to identify the location of each gastroscopy image based on a preset convolutional neural network model, and obtain the location category of the gastroscopy image; The second acquisition module is used to acquire N gastroscopy images belonging to each of the K parts belonging to the preset part category set, to obtain K*N gastroscopy images, where K and N are both positive integers; The classification module is used to classify each of the K*N gastroscopy images into P abnormal feature categories based on a preset large language model, and to obtain P abnormal classification confidence scores of the gastroscopy image belonging to the P abnormal feature categories, and to obtain P abnormal classification confidence scores of each of the K*N gastroscopy images, where P is a positive integer greater than 1. The determination module is used to determine the gastroscopy image classification category based on P abnormal classification confidence scores of each of the K*N gastroscopy images, wherein the gastroscopy image classification category is a normal category, a first abnormal image category, or a second abnormal image category.
[0011] In an optional embodiment, determining the gastroscopy image classification category based on P abnormality classification confidence scores for each of the K*N gastroscopy images includes: When N is greater than 1, each of the above-mentioned site categories is determined as the first site category, and each of the abnormal feature categories is determined as the first abnormal feature category, so as to obtain N abnormal classification confidence scores of N gastroscopy images under the first site category belonging to the first abnormal feature category; The average of the N abnormal classification confidence scores is determined as the feature confidence score of the gastroscopy image belonging to the first abnormal feature category under the first site category, thus obtaining K*P feature confidence scores of the gastroscopy image belonging to P abnormal feature categories under K site categories; The classification category of gastroscopy images is determined based on the confidence scores of K*P features.
[0012] In an optional embodiment, determining the gastroscopy image classification category based on K*P feature confidence scores includes: Obtain the baseline feature information of the target entity, which includes age and gender features; The baseline feature information of the target entity and the confidence scores of K*P features are input into a preset machine learning model to obtain the classification category of the gastroscopy image.
[0013] In an optional embodiment, the preset machine learning model is a random forest model, a decision tree model, or a support vector machine.
[0014] In an optional embodiment, K is 8, and the K site categories are respectively the gastric antrum category, duodenal bulb category, descending duodenal part category, lower part of the gastric body category (positive endoscope), upper and middle part of the gastric body category (positive endoscope), fundus category (reverse endoscope), upper and middle part of the gastric body category (reverse endoscope), and angle category (reverse endoscope).
[0015] In an optional embodiment, P is 8, and the P abnormal feature categories are old hemorrhage, redness, atrophy, fold swelling, goosebumps, mucosal swelling, intestinal metaplasia, and bile reflux.
[0016] Thirdly, the electronic device provided in this application includes a memory and a processor. The memory stores a computer program, and the processor runs the computer program in the memory to implement the steps in the information processing method for type A gastritis based on artificial intelligence provided in this application.
[0017] Fourthly, the computer-readable storage medium provided in this application stores multiple instructions that are adapted for loading by a processor to implement the steps in the information processing method for type A gastritis based on artificial intelligence provided in this application.
[0018] Fifthly, the computer program product provided in this application includes a computer program or instructions that, when executed by a processor, implement the steps in the information processing method for type A gastritis based on artificial intelligence provided in this application.
[0019] In this application, compared to related technologies, multiple gastroscopy images are acquired during a gastroscopy operation on a target entity; based on a preset convolutional neural network model, the gastroscopy images are used to identify the location of each location to obtain the location category; for each of the K location categories belonging to the preset location category set, N gastroscopy images belonging to the location category are acquired, resulting in K*N gastroscopy images, where K and N are both positive integers; based on a preset large language model, each of the K*N gastroscopy images is classified into P abnormal feature categories, resulting in P abnormal classification confidence scores for each of the P abnormal feature categories, and P abnormal classification confidence scores for each of the K*N gastroscopy images, where P is a positive integer greater than 1; based on the P abnormal classification confidence scores for each of the K*N gastroscopy images, the gastroscopy image classification category is determined, where the gastroscopy image classification category is a normal category, a first abnormal image category, or a second abnormal image category. This application can improve the accuracy of image classification. Attached Figure Description
[0020] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0021] Figure 1 This is a schematic diagram of a scenario for an artificial intelligence-based information processing system for type A gastritis provided in an embodiment of this application; Figure 2 This is a schematic flowchart of an embodiment of the information processing method for type A gastritis based on artificial intelligence provided in this application. Figure 3 This is a schematic diagram of another embodiment of the information processing method for type A gastritis based on artificial intelligence provided in this application; Figure 4 This is a schematic diagram of the structure of the artificial intelligence-based information processing device for type A gastritis provided in the embodiments of this application; Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation
[0022] It should be noted that the principles of this application are illustrated by example in a suitable computing environment. The following description is based on the specific embodiments of this application that are illustrated, and should not be regarded as limiting other specific embodiments not detailed herein.
[0023] In the following description of this application, "some embodiments" are referred to, which describe a subset of all possible embodiments. However, it is understood that "some embodiments" may be the same subset or different subset of all possible embodiments, and may be combined with each other without conflict.
[0024] In the following description of this application, the terms "first, second, third" are used merely to distinguish similar objects and do not represent a specific ordering of objects. It is understood that "first, second, third" may be interchanged in a specific order or sequence where permitted, so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein.
[0025] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.
[0026] To improve the effectiveness of information processing for type A gastritis based on artificial intelligence, embodiments of this application provide an information processing method, an information processing device, an electronic device, a computer-readable storage medium, and a computer program product for type A gastritis based on artificial intelligence. The information processing method for type A gastritis based on artificial intelligence can be executed by the information processing device for type A gastritis based on artificial intelligence, or by an electronic device integrating the information processing device for type A gastritis based on artificial intelligence.
[0027] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0028] This application also provides an AI-based information processing system for type A gastritis, which includes an electronic device. The electronic device integrates the AI-based information processing device for type A gastritis provided in this application.
[0029] To better understand the information processing method, apparatus, electronic device, and storage medium for type A gastritis based on artificial intelligence provided in the embodiments of this application, the application environment applicable to the embodiments of this application will be described below.
[0030] Please see Figure 1 , Figure 1 This diagram illustrates an application environment for an AI-based information processing method for type A gastritis according to an embodiment of this application. As one implementation, the AI-based information processing method for type A gastritis provided in this embodiment can be applied to an electronic device. This electronic device can be, for example,... Figure 1 The server 110 shown can be connected to the terminal device 120 via a network. The network serves as a medium for providing a communication link between the server 110 and the terminal device 120. The network can include various connection types, such as wired communication links, wireless communication links, etc., and this embodiment is not limited thereto. Optionally, in other embodiments, the electronic device can also be a smartphone, laptop, etc.
[0031] It should be understood that Figure 1 The server 110, network, and terminal device 120 shown are merely illustrative. Depending on the implementation requirements, any number of servers, networks, and terminal devices can be included. For example, server 110 can be a physical server or a server cluster consisting of multiple servers, and terminal device 120 can be a mobile phone, tablet, desktop computer, laptop computer, etc. It is understood that embodiments of this application can also allow multiple terminal devices 120 to access server 110 simultaneously.
[0032] In addition, the AI-based information processing system for type A gastritis may also include a memory for storing raw data, intermediate data, and result data during the AI-based information processing of type A gastritis. In this embodiment of the application, the storage device can be a cloud storage device. Cloud storage is a new concept that is extended and developed from the concept of cloud computing. A distributed cloud storage system (hereinafter referred to as a storage system) refers to a storage system that uses cluster applications, grid technology and distributed storage file system functions to bring together a large number of storage devices of various types in the network (storage devices are also called storage nodes) through application software or application interfaces to work together to provide data storage and business access functions to the outside world.
[0033] Currently, the storage method in storage systems is as follows: Logical volumes are created, and during creation, physical storage space is allocated to each logical volume. This physical storage space may consist of a single storage device or the disks of several storage devices. Clients store data on a logical volume, which means storing the data on the file system. The file system divides the data into many parts, each part being an object. Each object contains not only the data but also additional information such as a data identifier (ID entity). The file system writes each object to the physical storage space of that logical volume, and it records the storage location information of each object. Therefore, when a client requests access to data, the file system can allow the client to access the data based on the storage location information of each object.
[0034] The process by which a storage system allocates physical storage space to a logical volume is as follows: the physical storage space is pre-divided into strips according to the capacity estimate of the objects stored in the logical volume (this estimate often has a large margin relative to the actual capacity of the objects to be stored) and the grouping of Redundant Array of Independent Disks (RAID). A logical volume can be understood as a strip, thus allocating physical storage space to the logical volume.
[0035] It should be noted that, Figure 1 The schematic diagram of the AI-based information processing system for type A gastritis shown is merely an example. The AI-based information processing system and scenario for type A gastritis described in this application are intended to more clearly illustrate the technical solutions of this application and do not constitute a limitation on the technical solutions provided in this application. As those skilled in the art will know, with the evolution of AI-based information processing systems for type A gastritis and the emergence of new business scenarios, the technical solutions provided in this application are also applicable to similar technical problems.
[0036] The following sections provide detailed descriptions of each example. It should be noted that the sequence numbers of the following embodiments are not intended to limit the preferred order of the embodiments.
[0037] Please refer to Figure 2 , Figure 2 This is a schematic flowchart of an embodiment of the information processing method for type A gastritis based on artificial intelligence provided in this application. Figure 2 As shown, the flow of the information processing method for type A gastritis based on artificial intelligence provided in this application is as follows: 201. Obtain multiple gastroscopic images acquired during gastroscopy of the target entity.
[0038] In this embodiment of the application, the target entity can be a patient.
[0039] In one specific embodiment, when performing a gastroscopy on a target entity, image frames acquired by the endoscope are captured to obtain multiple gastroscopy images.
[0040] 202. Based on a pre-defined convolutional neural network model, perform site identification on each gastroscopy image to obtain the site category of the gastroscopy image.
[0041] In one specific embodiment, a pre-trained convolutional neural network model is used to identify the location of each gastroscopy image. Convolutional Neural Networks (CNNs) are a type of feedforward neural network with a deep structure that includes convolutional computations. They are one of the representative algorithms of deep learning. CNNs possess representation learning capabilities, enabling them to perform shift-invariant classification of input information according to their hierarchical structure; therefore, they are also known as "Shift-Invariant Artificial Neural Networks (SIANNs)".
[0042] The entire stomach body is divided into eight sub-regions based on its location: the antrum, duodenal bulb, descending duodenum, lower part of the stomach body (front view), upper and middle part of the stomach body (front view), fundus of the stomach (reverse view), upper and middle part of the stomach body (reverse view), and angle of the stomach (reverse view). A convolutional neural network is used to classify the gastroscopy images into nine region categories, which include the eight sub-regions and other regions.
[0043] In another specific embodiment, to improve prediction accuracy, site identification is performed on each gastroscopy image to obtain the site category of the gastroscopy image, including: (1) Each gastroscopy image is identified as the current gastroscopy image.
[0044] (2) The current gastroscopy image and the previous gastroscopy images are defined as multiple time-series images, which are arranged in order of their capture time.
[0045] (3) Input multiple time series images into a preset convolutional neural network model to obtain multiple first predicted part categories corresponding to multiple time series images.
[0046] (4) Input multiple time series images and corresponding multiple first predicted part categories into a preset long short memory model to obtain the target predicted part category.
[0047] The default Long Short-Term Memory (LSTM) model is used. The LSTM model outputs the body part category of the last image in a series of N consecutive images, which is also the target predicted body part category of the current gastroscopy image.
[0048] Furthermore, the first predicted site category of the current gastroscopy image output by the preset convolutional neural network model is obtained, and the second predicted site category of the current gastroscopy image output by the preset long short-term memory model is obtained. If the first predicted site category and the second predicted site category of the current gastroscopy image are the same, then the first predicted site category of the current gastroscopy image is determined as the target predicted site category of the current gastroscopy image; if the first predicted site category and the second predicted site category of the current gastroscopy image are different, then other sites are determined as the target predicted site categories of the current gastroscopy image.
[0049] 203. For each of the K part categories belonging to the preset part category set, obtain N gastroscopy images belonging to the part category, resulting in K*N gastroscopy images.
[0050] Where K and N are both positive integers. For example, K is 8 and N is 8, which can be set according to specific circumstances. The preset part category set includes 9 part categories, namely the gastric antrum, duodenal bulb, descending duodenum, lower part of the gastric body under positive endoscopy, upper and middle part of the gastric body under positive endoscopy, fundus of the gastric body under reverse endoscopy, upper and middle part of the gastric body under reverse endoscopy, angle of the gastric body under reverse endoscopy, and other parts.
[0051] K is 8, and the K site categories are: gastric antrum, duodenal bulb, descending duodenum, lower part of gastric body (positive endoscope), upper and middle part of gastric body (positive endoscope), fundus of gastric body (reverse endoscope), upper and middle part of gastric body (reverse endoscope), and angle of gastric body (reverse endoscope).
[0052] Specifically, for each of the K part categories belonging to the preset part category set, N gastroscopy images are obtained from multiple gastroscopy images of the part category.
[0053] In one specific embodiment, each body part category is designated as a second body part category. The capture timestamps of multiple gastroscopy images belonging to the second body part category are obtained. The earliest first time point and the latest second time point among these capture timestamps are then identified. The time interval between the first and second time points is divided into N equal time intervals. One gastroscopy image is captured within each time interval, resulting in N gastroscopy images for the second body part category. Alternatively, one gastroscopy image can be randomly selected from each time interval.
[0054] Furthermore, acquiring a gastroscopy image within each time interval includes: inputting each gastroscopy image within the time interval into a preset image segmentation model for segmentation to obtain multiple tissue segmentation regions on the gastroscopy image; determining the first area proportion of the tissue segmentation region belonging to the second part category on the gastroscopy image; obtaining the first area proportion of each gastroscopy image; and determining the gastroscopy image with the highest first area proportion within the time interval as the gastroscopy image acquired from the time interval.
[0055] Furthermore, the first area proportions of each gastroscopy image within the time interval are sorted from largest to smallest, and the top M gastroscopy images are obtained, where M is less than the total number of gastroscopy images within the time interval. The time interval between each of the M gastroscopy images and the median time point of the time interval is obtained, resulting in M time intervals corresponding to the M gastroscopy images. The gastroscopy image with the smallest time interval among the M gastroscopy images is determined as the gastroscopy image obtained from the time interval. Since the selected gastroscopy images are closer to the median time point, the gastroscopy images obtained from each time interval are more evenly distributed, thus providing a more comprehensive reflection of the stomach's condition.
[0056] 204. Based on a pre-defined large language model, classify each of the K*N gastroscopy images into P abnormal feature categories, and obtain the P abnormal classification confidence scores of the gastroscopy image belonging to the P abnormal feature categories.
[0057] Where P is a positive integer greater than 1.
[0058] Specifically, P is 8, and the P abnormal feature categories are old hemorrhage, redness, atrophy, swollen folds, goosebumps, mucosal swelling, intestinal metaplasia, and bile reflux. Of course, P can also be other values.
[0059] Specifically, the gastroscopy images are input into a pre-defined large language model to obtain the classification confidence scores for eight abnormal feature categories. The pre-defined large language model can classify P abnormal feature categories. The pre-defined large language model can be one of several models, such as chatGpt, claude, hunyuan, zhipu, moonshot, dashscope, or ark; the appropriate model can be selected based on the specific requirements.
[0060] To identify the image features of type A gastritis under gastroscopy, a large language model was used to extract the following eight features from the images: old bleeding, redness, atrophy, swollen folds, goosebump-like changes, mucosal swelling, intestinal metaplasia, and bile reflux. For a given input image, the abnormal classification confidence scores of the eight features were obtained.
[0061] For example, given a gastroscopy image, inputting the image into a pre-defined large language model yields P anomaly classification confidence scores for each of the P anomaly feature categories. Specifically, the gastroscopy image undergoes standardization preprocessing, including removing glare and adjusting the resolution to fit the model. A pre-trained image encoder extracts deep features such as texture and color, converting them into feature sequences. These sequences are then concatenated with text embedding vectors representing eight anomaly feature categories and input into the large language model. The model fuses text and image information through a cross-attention mechanism, and after multi-layer Transformer encoding and decoding, outputs a normalized probability value for each anomaly feature category—the eight anomaly classification confidence scores—with the sum of these values approaching 1.
[0062] For example, after inputting a gastroscopy image, the preset large language model outputs the following confidence levels for eight abnormal feature categories: old bleeding 0.02, redness 0.85, atrophy 0.01, fold swelling 0.03, goosebump-like changes 0.01, mucosal swelling 0.05, intestinal metaplasia 0.01, and bile reflux 0.02. This set of data intuitively reflects the probability of abnormality of the image in each category, among which the high confidence level of the redness category indicates that it should be given special attention.
[0063] 205. Determine the classification category of gastroscopy images based on the P abnormal classification confidence scores of each gastroscopy image in K*N gastroscopy images.
[0064] The gastroscopy images are classified into three categories: normal, first abnormal image, or second abnormal image. Specifically, the first abnormal image category is type A gastritis, and the second abnormal image category is non-type A gastritis.
[0065] In a specific embodiment, the gastroscopy image classification category is determined based on P abnormality classification confidence scores for each of the K*N gastroscopy images, including: (1) When N is greater than 1, each part category is determined as the first part category, and each abnormal feature category is determined as the first abnormal feature category, so as to obtain the N abnormal classification confidence scores of N gastroscopy images under the first part category belonging to the first abnormal feature category.
[0066] For example, the gastric antrum category is determined as the first location category, and the redness category is determined as the first abnormal feature category. For N gastroscopy images of the gastric antrum category, the abnormal classification confidence of each of the N gastroscopy images belonging to the redness category is calculated, thus obtaining the N abnormal classification confidences belonging to the first abnormal feature category.
[0067] (2) The average value of the N abnormal classification confidence scores is determined as the feature confidence score of the gastroscopy image belonging to the first abnormal feature category under the first location category, and K*P feature confidence scores of the gastroscopy image belonging to P abnormal feature categories under K location categories are obtained.
[0068] Since each body part category is designated as the first body part category, and each abnormal feature category is designated as the first abnormal feature category, each first body part category and each first abnormal feature category can obtain a feature confidence score. Therefore, K body part categories and P abnormal feature categories can yield K*P feature confidence scores. For example, if K is 8 and P is 8, then 8 body parts * 8 feature confidence scores = 64 feature confidence scores.
[0069] (3) Determine the classification category of gastroscopy images based on the confidence of K*P features.
[0070] In one specific embodiment, a pre-trained machine learning model is used, which may be a random forest model, a decision tree model, or a support vector machine. K*P feature confidence scores are input into the pre-trained machine learning model to obtain the classification category of the gastroscopy image.
[0071] In another specific embodiment, determining the classification category of the gastroscopy image based on K*P feature confidence scores includes: obtaining baseline feature information of the target entity, including age features, gender features, etc.; inputting the baseline feature information of the target entity and K*P feature confidence scores into a preset machine learning model to obtain the gastroscopy image classification category.
[0072] Please refer to Figure 3 , Figure 3 This is a schematic diagram of another embodiment of the information processing method for type A gastritis based on artificial intelligence provided in this application, as shown below. Figure 3 As shown, the flow of the information processing method for type A gastritis based on artificial intelligence provided in this application is as follows: (1) When performing gastroscopy on the target entity, capture video frames from the video captured by the endoscope to obtain multiple gastroscopy images.
[0073] (2) Based on the preset convolutional neural network model, the location of each gastroscopy image is identified to obtain the location category of the gastroscopy image.
[0074] (3) For each of the 8 part categories in the preset part category set, obtain N gastroscopy images belonging to the part category, and get 8*N gastroscopy images.
[0075] (4) Based on the preset large language model, classify each of the 8*N gastroscopy images into 8 abnormal feature categories, and obtain the 8 abnormal classification confidence scores of the gastroscopy images belonging to the 8 abnormal feature categories.
[0076] (5) When N is greater than 1, each part category is determined as the first part category, and each abnormal feature category is determined as the first abnormal feature category, so as to obtain the N abnormal classification confidence scores of N gastroscopy images under the first part category belonging to the first abnormal feature category.
[0077] (6) The average value of the N abnormal classification confidence scores is determined as the feature confidence score of the gastroscopy image belonging to the first abnormal feature category under the first location category, and the 8*8 feature confidence scores of the gastroscopy image belonging to the 8 abnormal feature categories under the 8 location categories are obtained.
[0078] (7) Input the baseline feature information of the target entity and the confidence scores of 8*8 features into the preset machine learning model to obtain the gastroscopy image classification category.
[0079] Specifically, the confidence scores of 8*8 features are used as feature values for 64 locations. The feature values of 64 locations and the baseline feature information of the target entity (age features, gender features, etc.) are input into a preset machine learning model to obtain the gastroscopy image classification category.
[0080] Among them, the gastroscopy images are classified into type A gastritis, non-type A gastritis, or normal.
[0081] To facilitate better implementation of the AI-based information processing method for type A gastritis provided in this application, this application also provides an AI-based information processing device for type A gastritis, based on the aforementioned AI-based information processing method. The meanings of the terms used are the same as in the aforementioned AI-based information processing method for type A gastritis; for specific implementation details, please refer to the descriptions in the above method embodiments.
[0082] Please refer to Figure 4 , Figure 4This is a schematic diagram of an embodiment of an AI-based information processing device for type A gastritis provided in this application. The AI-based information processing device for type A gastritis may include a first acquisition module 701, an identification module 702, a second acquisition module 703, a classification module 704, and a determination module 705. The first acquisition module 701 is used to acquire multiple gastroscopy images obtained when performing gastroscopy on the target entity; The recognition module 702 is used to identify the location of each gastroscopy image based on a preset convolutional neural network model, and obtain the location category of the gastroscopy image; The second acquisition module 703 is used to acquire N gastroscopy images belonging to each of the K part categories in the preset part category set, to obtain K*N gastroscopy images, where K and N are both positive integers; The classification module 704 is used to classify each of the K*N gastroscopy images into P abnormal feature categories based on a preset large language model, and obtain the P abnormal classification confidence scores of the gastroscopy image belonging to the P abnormal feature categories, and obtain the P abnormal classification confidence scores of each of the K*N gastroscopy images, where P is a positive integer greater than 1. The determination module 705 is used to determine the gastroscopy image classification category based on the P abnormal classification confidence scores of each gastroscopy image in K*N gastroscopy images, wherein the gastroscopy image classification category is a normal category, a first abnormal image category, or a second abnormal image category.
[0083] In an optional embodiment, the gastroscopy image classification category is determined based on P abnormality classification confidence scores for each of the K*N gastroscopy images, including: When N is greater than 1, each part category is determined as the first part category, and each abnormal feature category is determined as the first abnormal feature category. The N abnormal classification confidence scores of N gastroscopy images under the first part category belonging to the first abnormal feature category are obtained. The average of the N abnormality classification confidence scores is determined as the feature confidence score of the gastroscopy image belonging to the first abnormal feature category under the first location category, and K*P feature confidence scores of the gastroscopy image belonging to P abnormal feature categories under K location categories are obtained. The classification category of gastroscopy images is determined based on the confidence scores of K*P features.
[0084] In an optional embodiment, the classification category of the gastroscopy image is determined based on K*P feature confidence scores, including: Obtain baseline feature information of the target entity, including age and gender features; The baseline feature information of the target entity and the confidence scores of K*P features are input into a preset machine learning model to obtain the classification category of the gastroscopy image.
[0085] In an optional embodiment, the preset machine learning model is a random forest model, a decision tree model, or a support vector machine.
[0086] In an optional embodiment, K is 8, and the K site categories are the gastric antrum category, duodenal bulb category, descending duodenal part category, lower part of the gastric body category (positive endoscope), upper and middle part of the gastric body category (positive endoscope), fundus category (reverse endoscope), upper and middle part of the gastric body category (reverse endoscope), and angle category (reverse endoscope).
[0087] In an optional embodiment, P is 8, and the P abnormal feature categories are old hemorrhage, redness, atrophy, fold swelling, goosebumps, mucosal swelling, intestinal metaplasia, and bile reflux.
[0088] Please refer to Figure 5 , Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.
[0089] The electronic device may include a radio frequency (RF) circuit 901, a memory 902 including one or more computer-readable storage media, an input unit 903, a display unit 904, a sensor 905, an audio circuit 906, a wireless fidelity (WiFi) module 907, a processor 908 including one or more processing cores, and a power supply 909, among other components. Those skilled in the art will understand that... Figure 5 The electronic device structure shown does not constitute a limitation on the electronic device and may include more or fewer components than shown, or combine certain components, or have different component arrangements. Wherein: RF circuit 901 can be used for receiving and transmitting signals during information transmission or calls. Specifically, it receives downlink information from the base station and hands it over to one or more processors 908 for processing; additionally, it transmits uplink data to the base station. Typically, RF circuit 901 includes, but is not limited to, an antenna, at least one amplifier, a tuner, one or more oscillators, a Subscriber Identity Module (SIM) card, a transceiver, a coupler, a low-noise amplifier (LNA), a duplexer, etc. Furthermore, RF circuit 901 can also communicate wirelessly with networks and other devices. Wireless communication can use any communication standard or protocol, including but not limited to GSM, GPRS, CDMA, WCDMA, LTE, email, and SMS.
[0090] The memory 902 can be used to store software programs and modules. The processor 908 executes various functional applications and information processing based on artificial intelligence for type A gastritis by running the software programs and modules stored in the memory 902. The memory 902 may mainly include a program storage area and a data storage area. The program storage area may store the operating system, application programs required for at least one function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the electronic device (such as audio data, phone book, etc.). In addition, the memory 902 may include high-speed random access memory, and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other volatile solid-state storage device. Accordingly, the memory 902 may also include a memory controller to provide access to the memory 902 for the processor 908 and the input unit 903.
[0091] Input unit 903 can be used to receive input digital or character information, and to generate keyboard, mouse, joystick, optical, or trackball signal inputs related to user settings and function control. Specifically, in one embodiment, input unit 903 may include a touch-sensitive surface and other input devices. A touch-sensitive surface, also known as a touch display or touchpad, can collect user touch operations on or near it (e.g., user operations using fingers, styluses, or any suitable object or accessory on or near the touch-sensitive surface) and drive corresponding connection devices according to a pre-set program. Optionally, the touch-sensitive surface may include a touch detection device and a touch controller. The touch detection device detects the user's touch location and the signal generated by the touch operation, transmitting the signal to the touch controller; the touch controller receives touch information from the touch detection device, converts it into touch point coordinates, sends it to the processor 908, and can receive and execute commands from the processor 908. Furthermore, various types of touch-sensitive surfaces, such as resistive, capacitive, infrared, and surface acoustic wave, can be used. In addition to the touch-sensitive surface, input unit 903 may also include other input devices. Specifically, other input devices may include, but are not limited to, one or more of the following: physical keyboard, function keys (such as volume control buttons, power buttons, etc.), trackball, mouse, joystick, etc.
[0092] Display unit 904 can be used to display information input by the user or information provided to the user, as well as various graphical user interfaces of electronic devices. These graphical user interfaces can be composed of graphics, text, icons, video, and any combination thereof. Display unit 904 may include a display panel, optionally configured as a liquid crystal display (LCD), organic light-emitting diode (OLED), or similar form. Furthermore, a touch-sensitive surface may cover the display panel. When the touch-sensitive surface detects a touch operation on or near it, it transmits the information to processor 908 to determine the type of touch event. Subsequently, processor 908 provides corresponding visual output on the display panel according to the type of touch event. Although in the figures, the touch-sensitive surface and the display panel are shown as two separate components for implementing input and output functions, in some embodiments, the touch-sensitive surface and the display panel can be integrated to achieve both input and output functions.
[0093] Electronic devices may also include at least one sensor 905, such as a light sensor, a motion sensor, and other sensors. Specifically, the light sensor may include an ambient light sensor and a proximity sensor, wherein the ambient light sensor can adjust the brightness of the display panel according to the ambient light level, and the proximity sensor can turn off the display panel and / or backlight when the electronic device is moved to the ear. As a type of motion sensor, a gravity acceleration sensor can detect the magnitude of acceleration in various directions (generally three axes), and can detect the magnitude and direction of gravity when stationary. It can be used for applications that recognize the phone's posture (such as landscape / portrait switching, related games, magnetometer posture calibration), vibration recognition-related functions (such as pedometers, taps), etc. Other sensors that may be configured in electronic devices, such as gyroscopes, barometers, hygrometers, thermometers, and infrared sensors, will not be described in detail here.
[0094] Audio circuitry 906, a speaker, and a microphone provide an audio interface between the user and the electronic device. Audio circuitry 906 converts received audio data into electrical signals, transmits them to the speaker, and the speaker converts them into sound signals for output. Conversely, the microphone converts collected sound signals into electrical signals, which are then received by audio circuitry 906, converted back into audio data, and processed by processor 908. The processed data is then transmitted via RF circuitry 901 to, for example, another electronic device, or output to memory 902 for further processing. Audio circuitry 906 may also include an earphone jack to facilitate communication between external headphones and the electronic device.
[0095] WiFi is a short-range wireless transmission technology. Electronic devices using WiFi module 907 can help users send and receive emails, browse web pages, and access streaming media, providing users with wireless broadband internet access. Although WiFi module 907 is shown in the figure, it is understood that it is not an essential component of the electronic device and can be omitted as needed without changing the essence of the invention.
[0096] The processor 908 is the control center of the electronic device. It connects various parts of the phone via various interfaces and lines, and performs various functions and processes data by running or executing software programs and / or modules stored in the memory 902, and by calling data stored in the memory 902, thereby performing overall detection of the phone. Optionally, the processor 908 may include one or more processing cores; preferably, the processor 908 may integrate an application processor and a modem processor, wherein the application processor mainly handles the operating system, user interface, and applications, and the modem processor mainly handles wireless communication. It is understood that the modem processor may also not be integrated into the processor 908.
[0097] The electronic device also includes a power supply 909 (such as a battery) that supplies power to various components. Preferably, the power supply can be logically connected to the processor 908 through a power management system, thereby enabling functions such as charging, discharging, and power consumption management through the power management system. The power supply 909 may also include one or more DC or AC power supplies, recharging systems, power fault detection circuits, power converters or inverters, power status indicators, and other arbitrary components.
[0098] Although not shown, the electronic device may also include a camera, Bluetooth module, etc., which will not be described in detail here. Specifically, in this embodiment, the processor 908 in the electronic device loads the executable files corresponding to the processes of one or more applications into the memory 902 according to the following instructions, and the processor 908 runs the applications stored in the memory 902 to realize various functions: Compared to related technologies, this method acquires multiple gastroscopic images during a gastroscopy operation on a target entity; identifies the location of each gastroscopic image based on a pre-defined convolutional neural network model to obtain the location category of the gastroscopic image; for each of the K location categories belonging to the pre-defined location category set, acquires N gastroscopic images belonging to the location category, resulting in K*N gastroscopic images, where K and N are both positive integers; classifies each of the K*N gastroscopic images into P abnormal feature categories based on a pre-defined large language model, obtaining P abnormal classification confidence scores for each of the P abnormal feature categories, and obtaining P abnormal classification confidence scores for each of the K*N gastroscopic images, where P is a positive integer greater than 1; and determines the gastroscopic image classification category based on the P abnormal classification confidence scores for each of the K*N gastroscopic images, where the gastroscopic image classification category is either normal, a first abnormal image category, or a second abnormal image category.
[0099] It should be noted that the electronic device provided in this application embodiment and the information processing method for type A gastritis based on artificial intelligence in the above embodiment belong to the same concept. The specific implementation process can be found in the above related embodiments, and will not be repeated here.
[0100] This application also provides a computer-readable storage medium storing a computer program thereon. When the computer program stored thereon is executed on the processor of the electronic device provided in the embodiments of this application, the processor of the electronic device performs the steps in the information processing method for type A gastritis based on artificial intelligence provided in this application. The storage medium may be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.
[0101] This application also provides a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform various optional implementations of the aforementioned artificial intelligence-based information processing method for type A gastritis.
[0102] The above provides a detailed description of the information processing method and apparatus for type A gastritis based on artificial intelligence provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.
[0103] It should be noted that when the above embodiments of this application are applied to specific products or technologies, and user-related data is involved, user permission or consent is required, and the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions.
Claims
1. An information processing method for type A gastritis based on artificial intelligence, characterized in that, The information processing method for type A gastritis based on artificial intelligence includes: Acquire multiple gastroscopy images obtained during gastroscopy of the target entity; Based on a preset convolutional neural network model, the location of each gastroscopy image is identified to obtain the location category of the gastroscopy image; For each of the K location categories belonging to the preset location category set, N gastroscopic images belonging to the location category are obtained respectively, resulting in K*N gastroscopic images, where K and N are both positive integers; Based on a pre-defined large language model, each of the K*N gastroscopy images is classified into P abnormal feature categories, and P abnormal classification confidence scores of the gastroscopy image belonging to the P abnormal feature categories are obtained. P abnormal classification confidence scores of each of the K*N gastroscopy images are obtained, where P is a positive integer greater than 1. The gastroscopy image classification category is determined based on P abnormal classification confidence scores for each of the K*N gastroscopy images, wherein the gastroscopy image classification category is a normal category, a first abnormal image category, or a second abnormal image category.
2. The information processing method for type A gastritis based on artificial intelligence according to claim 1, characterized in that, The process of determining the gastroscopy image classification category based on P abnormality classification confidence scores for each of the K*N gastroscopy images includes: When N is greater than 1, each of the above-mentioned site categories is determined as the first site category, and each of the abnormal feature categories is determined as the first abnormal feature category, so as to obtain N abnormal classification confidence scores of N gastroscopy images under the first site category belonging to the first abnormal feature category; The average of the N abnormal classification confidence scores is determined as the feature confidence score of the gastroscopy image belonging to the first abnormal feature category under the first site category, thus obtaining K*P feature confidence scores of the gastroscopy image belonging to P abnormal feature categories under K site categories; The classification category of gastroscopy images is determined based on the confidence scores of K*P features.
3. The information processing method for type A gastritis based on artificial intelligence according to claim 2, characterized in that, The method of determining the classification category of gastroscopy images based on K*P feature confidence scores includes: Obtain the baseline feature information of the target entity, which includes age and gender features; The baseline feature information of the target entity and the confidence scores of K*P features are input into a preset machine learning model to obtain the classification category of the gastroscopy image.
4. The information processing method for type A gastritis based on artificial intelligence according to claim 3, characterized in that, The preset machine learning model is a random forest model, a decision tree model, or a support vector machine.
5. The information processing method for type A gastritis based on artificial intelligence according to claim 3, characterized in that, K is 8, and the K categories of the described parts are: gastric antrum, duodenal bulb, descending duodenum, lower part of the gastric body under normal vision, upper and middle part of the gastric body under normal vision, fundus of the gastric body under normal vision, upper and middle part of the gastric body under normal vision, and angle of the gastric body under normal vision.
6. The information processing method for type A gastritis based on artificial intelligence according to claim 3, characterized in that, P is 8, and the P abnormal feature categories are old hemorrhage, redness, atrophy, fold swelling, goosebumps, mucosal swelling, intestinal metaplasia, and bile reflux.
7. An information processing device for type A gastritis based on artificial intelligence, characterized in that, The artificial intelligence-based information processing device for type A gastritis includes: The first acquisition module is used to acquire multiple gastroscopy images obtained when performing gastroscopy on the target entity; The recognition module is used to identify the location of each gastroscopy image based on a preset convolutional neural network model, and obtain the location category of the gastroscopy image; The second acquisition module is used to acquire N gastroscopy images belonging to each of the K parts belonging to the preset part category set, to obtain K*N gastroscopy images, where K and N are both positive integers; The classification module is used to classify each of the K*N gastroscopy images into P abnormal feature categories based on a preset large language model, and to obtain P abnormal classification confidence scores of the gastroscopy image belonging to the P abnormal feature categories, and to obtain P abnormal classification confidence scores of each of the K*N gastroscopy images, where P is a positive integer greater than 1. The determination module is used to determine the gastroscopy image classification category based on P abnormal classification confidence scores of each of the K*N gastroscopy images, wherein the gastroscopy image classification category is a normal category, a first abnormal image category, or a second abnormal image category.
8. An electronic device, characterized in that, The method includes a memory and a processor, the memory storing a computer program, and the processor running the computer program in the memory to perform the steps of the information processing method for type A gastritis based on artificial intelligence as described in any one of claims 1 to 6.
9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a plurality of instructions adapted for loading by a processor to perform the steps of the information processing method for type A gastritis based on artificial intelligence as described in any one of claims 1 to 6.
10. A computer program product, comprising a computer program or instructions, characterized in that, When the computer program or instructions are executed by a processor, they implement the steps in the information processing method for type A gastritis based on artificial intelligence as described in any one of claims 1 to 6.