Gastrointestinal endoscope intelligent detection system and method

Abstract

The invention provides a gastrointestinal endoscope intelligent detection system and method, and the system comprises an image collection module which is used for collecting gastrointestinal endoscope images, and generating a high-resolution video stream or image sequence; the image data set construction module is used for constructing a gastrointestinal endoscope image data set and labeling intestinal wall lesion tissues in the image to form a training data set; the lesion tissue intelligent identification module is used for automatically identifying intestinal wall lesion tissues in the gastrointestinal endoscope image and carrying out three-dimensional positioning and size measurement on the intestinal wall lesion tissues; and the automatic alarm module is used for automatically sending out an alarm signal when the diseased tissue is identified. According to the technical scheme, the recognition rate and the positioning precision of the intestinal wall diseased tissue are improved, missed diagnosis and misdiagnosis are reduced, and the treatment effect is improved.

Description

Background Art

[0002] With the increasing incidence of digestive tract diseases year by year, gastroscopy has become an important means for diagnosing and treating digestive tract diseases. However, traditional gastroscopy highly relies on doctors' experience and naked-eye observation, which is prone to missed diagnosis and misdiagnosis. Especially, it is difficult to identify and locate tiny lesions such as small polyps. In addition, during the process of removing or biopsying diseased tissues, the operation precision and speed of doctors also directly affect the treatment effect. Summary of the Invention

[0003] This application provides a gastroscope intelligent detection system and method to improve the recognition rate and positioning accuracy of intestinal wall diseased tissues, reduce missed diagnosis and misdiagnosis, and improve the treatment effect.

[0004] In the first aspect, a gastroscope intelligent detection system is provided, including:

[0005] An image acquisition module, configured to acquire gastroscope images and generate high-resolution video streams or image sequences;

[0006] An image data set construction module, configured to construct a gastroscope image data set and label the intestinal wall diseased tissues in the images to form a training data set;

[0007] A diseased tissue intelligent recognition module, configured to automatically recognize the intestinal wall diseased tissues in the gastroscope images and perform three-dimensional positioning and size measurement on the intestinal wall diseased tissues;

[0008] An automatic alarm module, configured to automatically emit an alarm signal when intestinal wall diseased tissues are recognized.

[0009] In the above technical solution, by setting an image acquisition module, configured to acquire gastroscope images and generate high-resolution video streams or image sequences; an image data set construction module, configured to construct a gastroscope image data set and label the intestinal wall diseased tissues in the images to form a training data set; a diseased tissue intelligent recognition module, configured to automatically recognize the intestinal wall diseased tissues in the gastroscope images and perform three-dimensional positioning and size measurement on the intestinal wall diseased tissues; an automatic alarm module, configured to automatically emit an alarm signal when intestinal wall diseased tissues are recognized; by constructing a gastroscope image data set, intelligently recognizing the three-dimensional positioning and size of intestinal wall diseased tissues, and integrating an automatic alarm, the recognition rate and positioning accuracy of intestinal wall diseased tissues are significantly improved, missed diagnosis and misdiagnosis are reduced, and the treatment effect is improved.

[0010] In a specific feasible implementation, it further includes:

[0011] An auxiliary treatment module, configured to provide real-time navigation and guidance according to the three-dimensional positioning information of the intestinal wall diseased tissues to assist doctors in removing or biopsying the intestinal wall diseased tissues.

[0012] In a specific feasible implementation, it further includes:

[0013] A data storage and management module for storing and managing gastroscopy image data, lesion recognition results, and treatment operation record information.

[0014] In a specific feasible implementation, based on a deep learning algorithm, the intestinal wall lesion tissues in gastroscopy images are automatically recognized.

[0015] In a specific feasible implementation, the steps of three-dimensional positioning of intestinal wall lesion tissues specifically include:

[0016] Using multi-view gastroscopy images, calculating the three-dimensional coordinates of intestinal wall lesion tissues through a stereo matching algorithm;

[0017] According to the three-dimensional coordinates of the intestinal wall lesion tissues, combined with the digestive tract anatomical structure, three-dimensional reconstruction of the intestinal wall lesion tissues is performed to generate a three-dimensional model of the intestinal wall lesion tissues;

[0018] Using the three-dimensional model of the intestinal wall lesion tissues to perform three-dimensional positioning of the intestinal wall lesion tissues.

[0019] In a specific feasible implementation, the steps of providing real-time navigation and guidance according to the three-dimensional positioning information of intestinal wall lesion tissues specifically include:

[0020] Overlaying a virtual scale on the gastroscopy image to display the distance and orientation between the lesion and the end of the endoscope;

[0021] Automatically calculating the optimal operation path according to the lesion position and the current posture of the endoscope.

[0022] In a specific feasible implementation, the alarm methods of the automatic alarm module include:

[0023] Visual alarm, which is used to highlight the lesion position on the display screen and pop up an alarm message;

[0024] Sound alarm, which is used to give an alarm through a speaker.

[0025] In a second aspect, a gastroscopy intelligent detection method is provided, including the following steps:

[0026] Using an image acquisition module to acquire gastroscopy images and generate a high-resolution video stream or image sequence;

[0027] Using an image data set construction module to construct a gastroscopy image data set and annotate the intestinal wall lesion tissues in the images to form a training data set;

[0028] The intelligent recognition module for diseased tissues automatically recognizes the intestinal wall diseased tissues in the gastroscope images, and performs three-dimensional positioning and size measurement on the intestinal wall diseased tissues;

[0029] The automatic alarm module automatically emits an alarm signal when intestinal wall diseased tissues are recognized.

[0030] In the above technical solution, an image acquisition module is set up to acquire gastroscope images and generate high-resolution video streams or image sequences; an image dataset construction module is used to construct a gastroscope image dataset and label the intestinal wall diseased tissues in the images to form a training dataset; the intelligent recognition module for diseased tissues is used to automatically recognize the intestinal wall diseased tissues in the gastroscope images and perform three-dimensional positioning and size measurement on the intestinal wall diseased tissues; the automatic alarm module is used to automatically emit an alarm signal when intestinal wall diseased tissues are recognized; by constructing a gastroscope image dataset, intelligently recognizing the three-dimensional positioning and size of intestinal wall diseased tissues, and integrating an automatic alarm, the recognition rate and positioning accuracy of intestinal wall diseased tissues are significantly improved, reducing missed diagnosis and misdiagnosis and improving the treatment effect.

[0031] In a specific feasible implementation, it further includes:

[0032] The auxiliary treatment module provides real-time navigation and guidance according to the three-dimensional positioning information of the intestinal wall diseased tissues to assist the doctor in removing or biopsying the intestinal wall diseased tissues.

[0033] In a specific feasible implementation, it further includes:

[0034] The data storage and management module stores and manages gastroscope image data, lesion recognition results, and treatment operation record information. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 It is a structural block diagram of the gastroscope intelligent detection system provided by an embodiment of the present application;

[0036] Figure 2 It is a flow block diagram of the gastroscope intelligent detection method provided by an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0037] The present application will be further described in detail below with reference to the drawings and embodiments. Through these descriptions, the features and advantages of the present application will become more clearly defined.

[0038] The special word "exemplary" here means "serving as an example, embodiment, or illustrative". Any embodiment described here as "exemplary" does not have to be interpreted as superior to or better than other embodiments. Although various aspects of the embodiments are shown in the drawings, the drawings do not have to be drawn to scale unless otherwise specified.

[0039] In addition, the technical features involved in different embodiments of the present application described below can be combined with each other as long as they do not conflict with each other.

[0040] To facilitate the understanding of the gastroscope intelligent detection system and method provided by the embodiments of the present application, its application scenarios will be described first. The gastroscope intelligent detection system and method provided by the embodiments of the present application are used to improve the recognition rate and positioning accuracy of intestinal wall lesion tissues, reduce missed diagnoses and misdiagnoses, and improve the treatment effect. With the increasing annual incidence of digestive tract diseases, gastroscope examination has become an important means for diagnosing and treating digestive tract diseases. However, traditional gastroscope examination highly relies on doctors' experience and naked-eye observation, and is prone to missed diagnoses and misdiagnoses. In particular, it is difficult to identify and locate tiny lesions such as small polyps. In addition, during the process of removing or biopsying lesion tissues, the operation precision and speed of doctors also directly affect the treatment effect. For this reason, the embodiments of the present application provide a gastroscope intelligent detection system and method to improve the recognition rate and positioning accuracy of intestinal wall lesion tissues, reduce missed diagnoses and misdiagnoses, and improve the treatment effect. The following will be described in detail with specific drawings by way of examples.

[0041] Reference Figure 1 and Figure 2 , Figure 1 is the structural block diagram of the gastroscope intelligent detection system provided by the embodiment of the present application; Figure 2 is the flow block diagram of the gastroscope intelligent detection method provided by the embodiment of the present application.

[0042] In Figure 1 , the embodiment of the present application provides a gastroscope intelligent detection system, including:

[0043] An image acquisition module, configured to acquire gastroscope images and generate a high-resolution video stream or image sequence;

[0044] An image data set construction module, configured to construct a gastroscope image data set and label the intestinal wall lesion tissues in the images to form a training data set;

[0045] A lesion tissue intelligent recognition module, configured to automatically recognize the intestinal wall lesion tissues in the gastroscope images, and perform three-dimensional positioning and size measurement on the intestinal wall lesion tissues;

[0046] An automatic alarm module, configured to automatically send an alarm signal when an intestinal wall lesion tissue is recognized.

[0047] In the above technical solution, an image acquisition module is provided for acquiring gastroscope images and generating high-resolution video streams or image sequences; an image dataset construction module is used to construct a gastroscope image dataset and label the intestinal wall lesion tissues in the images to form a training dataset; a lesion tissue intelligent recognition module is used to automatically recognize the intestinal wall lesion tissues in the gastroscope images and perform three-dimensional positioning and size measurement on the intestinal wall lesion tissues; an automatic alarm module is used to automatically send an alarm signal when an intestinal wall lesion tissue is recognized; by constructing a gastroscope image dataset, intelligently recognizing the three-dimensional positioning and size of intestinal wall lesion tissues, and integrating an automatic alarm, the recognition rate and positioning accuracy of intestinal wall lesion tissues are significantly improved, reducing missed diagnoses and misdiagnoses and improving the treatment effect.

[0048] In a specific feasible implementation scheme, it further includes:

[0049] An auxiliary treatment module is used to provide real-time navigation and guidance according to the three-dimensional positioning information of the intestinal wall lesion tissues to assist the doctor in removing or biopsying the intestinal wall lesion tissues.

[0050] In a specific feasible implementation scheme, it further includes:

[0051] A data storage and management module is used to store and manage gastroscope image data, lesion recognition results, and treatment operation record information.

[0052] In a specific feasible implementation scheme, based on a deep learning algorithm, the intestinal wall lesion tissues in the gastroscope images are automatically recognized.

[0053] In a specific feasible implementation scheme, the steps of performing three-dimensional positioning on the intestinal wall lesion tissues specifically include:

[0054] Using multi-view gastroscope images, calculating the three-dimensional coordinates of the intestinal wall lesion tissues through a stereo matching algorithm;

[0055] According to the three-dimensional coordinates of the intestinal wall lesion tissues and combining with the digestive tract anatomical structure, performing three-dimensional reconstruction on the intestinal wall lesion tissues to generate a three-dimensional model of the intestinal wall lesion tissues;

[0056] Using the three-dimensional model of the intestinal wall lesion tissues to perform three-dimensional positioning on the intestinal wall lesion tissues.

[0057] In a specific feasible implementation scheme, the steps of providing real-time navigation and guidance according to the three-dimensional positioning information of the intestinal wall lesion tissues specifically include:

[0058] Overlaying a virtual scale on the gastroscope image to display the distance and orientation between the lesion and the end of the endoscope;

[0059] Automatically calculating the optimal operation path according to the lesion position and the current posture of the endoscope.

[0060] In a specific feasible embodiment, the alarm methods of the automatic alarm module include:

[0061] Visual alarm, which is used to highlight the lesion location on the display screen and pop up an alarm message;

[0062] Sound alarm, which is used to give an alarm through a speaker.

[0063] Specifically, the gastrointestinal endoscope intelligent detection system includes:

[0064] Image acquisition module: It is used to acquire gastrointestinal endoscope images inside the patient's digestive tract and generate high-resolution video streams or image sequences.

[0065] Image dataset construction module: It is used to construct a gastrointestinal endoscope image dataset containing normal tissues and diseased tissues, and label the intestinal wall diseased tissues in the images through a data annotation tool to form a training dataset.

[0066] Diseased tissue intelligent recognition module: Based on deep learning algorithms, it automatically recognizes the intestinal wall diseased tissues in gastrointestinal endoscope images and realizes three-dimensional positioning and size measurement of the lesions.

[0067] Automatic alarm module: When intestinal wall diseased tissues are recognized, the system can automatically send out an alarm signal to remind the doctor to conduct further examinations and treatments.

[0068] Auxiliary removal module: According to the three-dimensional positioning information of the intestinal wall diseased tissues, the system provides real-time navigation and guidance to assist the doctor in accurately removing the intestinal wall diseased tissues.

[0069] Data storage and management module: It is used to store and manage information such as the patient's gastrointestinal endoscope image data, lesion recognition results, and removal operation records, which is convenient for subsequent query and analysis.

[0070] Furthermore, the construction of the image dataset includes:

[0071] Data acquisition: Obtain gastrointestinal endoscope images inside the patient's digestive tract through the image acquisition module. The image data includes high-resolution static images and dynamic video streams. To improve the diversity and representativeness of the dataset, the collected images cover patient data of different ages, genders, lesion types, and lesion sites.

[0072] Data annotation: Use a data annotation tool to manually annotate the collected gastrointestinal endoscope images. The annotation content includes the location, size, and shape information of the intestinal wall diseased tissues. The annotated dataset is used to train a deep learning model to improve the recognition accuracy and generalization ability of the model.

[0073] Furthermore, the intelligent recognition includes:

[0074] 1. Deep learning model: A deep learning model using a convolutional neural network (CNN) automatically identifies intestinal wall lesion tissues in gastroscopy images. The input of the model is the gastroscopy image, and the output is the location, size, and confidence score of the lesion.

[0075] 2. Three-dimensional positioning: Through multi-view image fusion and three-dimensional reconstruction technology, three-dimensional positioning of intestinal wall lesion tissues is achieved. The specific methods include:

[0076] Using multi-view gastroscopy images, calculate the three-dimensional coordinates of the lesion through a stereo matching algorithm;

[0077] Combined with the digestive tract anatomical structure, perform three-dimensional reconstruction on the lesion to generate a three-dimensional model of the lesion.

[0078] 3. Size recognition: Based on the three-dimensional positioning results, calculate the size information of intestinal wall lesion tissues, including parameters such as length, width, height, and volume. The size recognition results are used to guide subsequent removal operations.

[0079] Furthermore, the automatic alarm includes:

[0080] When the lesion tissue intelligent recognition module detects a lesion, the system automatically calculates the confidence score and size information of the lesion. If the confidence score of the lesion exceeds the preset threshold and the size information meets the diagnostic criteria for lesion tissues, the system immediately issues an alarm signal to remind the doctor to conduct further examinations and treatments.

[0081] The alarm methods include:

[0082] Visual alarm: Highlight the lesion location on the display screen and pop up an alarm message.

[0083] Sound alarm: Alarm through the speaker.

[0084] Furthermore, the auxiliary removal module includes:

[0085] 1. Real-time navigation unit

[0086] Based on the three-dimensional positioning information of intestinal wall lesion tissues, the system provides a real-time navigation function to help doctors quickly locate the lesion location. The specific implementation methods include:

[0087] Virtual scale: Overlay a virtual scale on the gastroscopy image to display the distance and orientation between the lesion and the end of the endoscope, helping doctors accurately judge the operation path.

[0088] Path planning: Automatically calculate the optimal operation path according to the lesion location and the current posture of the endoscope, and display it to the doctor through a graphical interface to reduce operation errors.

[0089] 2. Precision Excision Unit

[0090] Under the guidance of navigation information, doctors can use endoscopic instruments to perform excision operations on diseased tissues. To improve the precision and safety of excision, the system provides the following auxiliary functions:

[0091] Instrument Tracking: Through image processing technology, the position and posture of the endoscopic instrument are tracked in real time to ensure that the instrument is always in the optimal operating position.

[0092] Force Feedback: Combined with a force sensor, the force of the excision operation is monitored in real time to avoid tissue damage caused by excessive force.

[0093] Operation Record: The system automatically records the whole process of the excision operation, including information such as operation time, force, and path, which is convenient for subsequent analysis and evaluation.

[0094] Furthermore, data storage and management include:

[0095] 1. Data Storage

[0096] The system stores information such as the gastroscope and colonoscope image data of the patient, the lesion recognition results, and the excision operation records in the database for subsequent query and analysis. Encryption technology is used for data storage to ensure patient privacy and data security.

[0097] 2. Data Management

[0098] The system provides data management functions, including data query, data analysis, data export, etc. Doctors can query historical data according to conditions such as patient ID, examination date, and lesion type, and generate statistical analysis reports to assist clinical decision-making.

[0099] Furthermore, the gastroscope and colonoscope intelligent detection system provided by the present invention can be integrated into existing gastroscope and colonoscope examination equipment, and communicate with devices such as the endoscopic host, image processor, and display through an interface protocol to achieve seamless integration of functions.

[0100] In the above technical solution, the recognition rate and positioning accuracy of intestinal wall diseased tissues can be significantly improved, reducing missed diagnosis and misdiagnosis and improving the treatment effect. The automatic alarm and auxiliary excision functions of the system effectively reduce the doctor's workload and improve the safety and efficiency of the operation.

[0101] In a specific feasible implementation, the gastroscope and colonoscope intelligent detection system includes:

[0102] Image Acquisition: Use high-resolution gastroscope and colonoscope equipment to collect image data of the patient's digestive tract to generate a video stream and static images.

[0103] Data annotation: Manually annotate the collected images to mark the location and size of the intestinal wall lesion tissues, and construct a training dataset.

[0104] Model training: Use the annotated dataset to train a deep learning model, and optimize the recognition accuracy and generalization ability of the model.

[0105] Lesion recognition: Apply the trained model to new gastroscopy images, automatically recognize the intestinal wall lesion tissues, and perform three-dimensional positioning and size measurement.

[0106] Automatic alarm: When intestinal wall lesion tissues are recognized, the system emits an alarm signal to remind the doctor to conduct further examinations and treatments.

[0107] Assisted removal: According to the three-dimensional positioning information of the lesion, the system provides real-time navigation and guidance to assist the doctor in accurately removing the lesion tissues.

[0108] Data storage: Store information such as the patient's image data, lesion recognition results, and removal operation records in a database for subsequent query and analysis.

[0109] In a specific feasible implementation, the gastroscope intelligent detection system includes:

[0110] Multi-view fusion module: Use multi-view gastroscope images to calculate the three-dimensional coordinates of the lesion through a stereo matching algorithm to achieve three-dimensional positioning of the lesion.

[0111] Three-dimensional reconstruction module: Combine the digestive tract anatomical structure to perform three-dimensional reconstruction on the lesion, generate a three-dimensional model of the lesion, and provide more intuitive visual information.

[0112] Real-time navigation module: Overlay a virtual scale on the gastroscope image to display the lesion location and operation path, and help the doctor accurately judge the operation path.

[0113] Force feedback module: Combine a force sensor to monitor the force of the removal operation in real time to avoid tissue damage caused by excessive force.

[0114] Operation record module: The system automatically records the whole process of the removal operation, including information such as operation time, force, and path, for subsequent analysis and evaluation.

[0115] Furthermore, the gastroscope intelligent detection system includes:

[0116] Transfer learning module: Use a pre-trained model for transfer learning to reduce the training time and improve the generalization ability of the model. By pre-training on a large-scale medical image dataset and then fine-tuning on a specific gastroscope image dataset, the recognition accuracy of the model can be significantly improved.

[0117] Multi-task learning module: Integrate tasks such as lesion recognition, three-dimensional localization, and size measurement into a multi-task learning framework. By sharing feature representations, improve the performance of each task.

[0118] Model compression module: Utilize model compression techniques such as pruning, quantization, knowledge distillation, etc. to reduce the volume of the model, improve the running speed of the model, and meet the requirements of real-time detection.

[0119] Furthermore, the gastroscope intelligent detection system includes:

[0120] Multi-center data acquisition module: Collect gastroscope image data from multiple medical centers, covering data under different populations, different devices, and different examination conditions to increase the diversity of data.

[0121] Data augmentation module: Use data augmentation techniques such as rotation, scaling, flipping, adding noise, etc. to generate more training samples and increase the richness of the dataset.

[0122] Annotation optimization module: Adopt semi-automatic annotation tools and combine manual verification to improve the efficiency and accuracy of annotation.

[0123] Furthermore, the gastroscope intelligent detection system includes:

[0124] Multi-lesion detection module: Expand the system function to enable it to detect and identify various digestive tract lesions, such as ulcers, tumors, inflammation, etc.

[0125] Multi-device compatibility module: Optimize the system architecture to enable it to be compatible with gastroscope devices of different brands and models, and improve the universality of the system.

[0126] Remote diagnosis module: Combine 5G technology and cloud computing platform to achieve remote gastroscope examination and diagnosis, and improve the utilization efficiency of medical resources.

[0127] In Figure 2 this application embodiment provides a gastroscope intelligent detection method, including the following steps:

[0128] Use the image acquisition module to collect gastroscope images and generate high-resolution video streams or image sequences;

[0129] Use the image dataset construction module to construct a gastroscope image dataset, and annotate the intestinal wall lesion tissues in the images to form a training dataset;

[0130] Use the intelligent lesion tissue recognition module to automatically identify the intestinal wall lesion tissues in the gastroscope images, and perform three-dimensional localization and size measurement on the intestinal wall lesion tissues;

[0131] Use the automatic alarm module to automatically send an alarm signal when intestinal wall lesion tissues are recognized.

[0132] In the above technical solution, an image acquisition module is provided for acquiring gastroscope images and generating a high-resolution video stream or image sequence; an image dataset construction module is used to construct a gastroscope image dataset and label the intestinal wall lesion tissues in the images to form a training dataset; a lesion tissue intelligent recognition module is used to automatically recognize the intestinal wall lesion tissues in the gastroscope images and perform three-dimensional positioning and size measurement on the intestinal wall lesion tissues; an automatic alarm module is used to automatically send an alarm signal when intestinal wall lesion tissues are recognized. By constructing a gastroscope image dataset, intelligently recognizing the three-dimensional positioning and size of intestinal wall lesion tissues, and integrating an automatic alarm, the recognition rate and positioning accuracy of intestinal wall lesion tissues are significantly improved, reducing missed diagnosis and misdiagnosis and improving the treatment effect.

[0133] In a specific feasible implementation, it further includes:

[0134] The auxiliary treatment module is used to provide real-time navigation and guidance according to the three-dimensional positioning information of the intestinal wall lesion tissues to assist the doctor in removing or biopsying the lesion tissues.

[0135] In a specific feasible implementation, it further includes:

[0136] The data storage and management module is used to store and manage gastroscope image data, lesion recognition results, and treatment operation record information.

[0137] Those skilled in the art of the present technology know that the present application can be implemented as a system, a method, or a computer program product.

[0138] Therefore, the present disclosure can be specifically implemented in the following forms, that is: it can be completely hardware, can also be completely software (including firmware, resident software, microcode, etc.), and can also be a combination of hardware and software, generally referred to as "circuit", "module", or "system" in this article. In addition, in some embodiments, the present application can also be implemented in the form of a computer program product in one or more computer-readable media, which contains computer-readable program code.

[0139] Any combination of one or more computer-readable media may be employed. The computer-readable media may be a computer-readable signal medium or a computer-readable storage medium. A computer-readable storage medium may be, for example — but not limited to — an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples (a non-exhaustive list) of the computer-readable storage medium include: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the present document, a computer-readable storage medium may be any tangible medium that contains or stores a program which can be used by or in connection with an instruction execution system, apparatus, or device.

[0140] Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present application. Those of ordinary skill in the art may make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present application. On this basis, various substitutions and improvements can be made to the present application, and all of these fall within the protection scope of the present application.

Claims

1. An intelligent gastroscope detection system, characterized in that, Including: An image acquisition module, which is used to acquire gastroscope images and generate a high-resolution video stream or image sequence; An image dataset construction module, which is used to construct a gastroscope image dataset, label the intestinal wall lesion tissues in the images, and form a training dataset; A lesion tissue intelligent recognition module, which is used to automatically recognize the intestinal wall lesion tissues in the gastroscope images, and perform three-dimensional positioning and size measurement on the intestinal wall lesion tissues; An automatic alarm module, which is used to automatically send an alarm signal when intestinal wall lesion tissues are recognized.

2. The gastrointestinal endoscope intelligent detection system according to claim 1, wherein, It also includes: An auxiliary treatment module, which is used to provide real-time navigation and guidance according to the three-dimensional positioning information of the intestinal wall lesion tissues, and assist doctors in removing or biopsying the intestinal wall lesion tissues.

3. The gastrointestinal endoscope intelligent detection system according to claim 2, characterized in that, It also includes: A data storage and management module, which is used to store and manage gastroscope image data, lesion recognition results, and treatment operation record information.

4. The gastrointestinal endoscope intelligent detection system according to claim 3, characterized in that, Based on deep learning algorithms, automatically recognize the intestinal wall lesion tissues in the gastroscope images.

5. The gastrointestinal endoscope intelligent detection system according to claim 4, characterized in that, The steps for three-dimensional positioning of the intestinal wall lesion tissues specifically include: Using multi-view gastroscope images, calculating the three-dimensional coordinates of the intestinal wall lesion tissues through a stereo matching algorithm; According to the three-dimensional coordinates of the intestinal wall lesion tissues, combining with the digestive tract anatomical structure, performing three-dimensional reconstruction on the intestinal wall lesion tissues to generate a three-dimensional model of the intestinal wall lesion tissues; Using the three-dimensional model of the intestinal wall lesion tissues to perform three-dimensional positioning on the intestinal wall lesion tissues.

6. The gastrointestinal endoscope intelligent detection system according to claim 5, characterized in that, The steps for providing real-time navigation and guidance according to the three-dimensional positioning information of the intestinal wall lesion tissues specifically include: Overlaying a virtual scale on the gastroscope image to display the distance and orientation between the lesion and the end of the endoscope; Automatically calculating the optimal operation path according to the lesion position and the current posture of the endoscope.

7. The gastrointestinal endoscope intelligent detection system according to claim 6, wherein, The alarm methods of the automatic alarm module include: Visual alarm, which is used to highlight the lesion position on the display screen and pop up an alarm message; Sound alarm, which is used to give an alarm through a speaker.

8. An intelligent detection method for gastroscopy and colonoscopy, characterized in that, Including the following steps: Using the image acquisition module to acquire gastroscope images and generate a high-resolution video stream or image sequence; Using the image dataset construction module to construct a gastroscope image dataset, label the intestinal wall lesion tissues in the images, and form a training dataset; Using the lesion tissue intelligent recognition module to automatically recognize the intestinal wall lesion tissues in the gastroscope images, and perform three-dimensional positioning and size measurement on the intestinal wall lesion tissues; Using the automatic alarm module to automatically send an alarm signal when intestinal wall lesion tissues are recognized.

9. The gastroscope intelligent detection method according to claim 8, wherein, It also includes: Using the auxiliary treatment module to provide real-time navigation and guidance according to the three-dimensional positioning information of the intestinal wall lesion tissues, and assist doctors in removing or biopsying the intestinal wall lesion tissues.

10. The gastrointestinal endoscope intelligent detection method according to claim 9, wherein, It also includes: Using the data storage and management module to store and manage gastroscope image data, lesion recognition results, and treatment operation record information.

Images