Intelligent gastrointestinal endoscope operation auxiliary system and method
Through the intelligent gastroenteroscopy operation assistance system, the gastroenteroscopy path is optimized using multi-objective genetic algorithm and neural network model, and the problem of traditional manual path planning is solved, achieving efficient and accurate examination path planning and patient recovery Evaluation improves examination efficiency and treatment effectiveness.
Patent Information
- Application Number
- CN202411121406.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-15
- Publication Date
- 2024-11-26
AI Technical Summary
Manually planning the examination path of traditional gastroenteroscopy is difficult and time-consuming, which can easily lead to misoperation, increase the risk of patient discomfort and missed diagnosis, and cannot fully cover the lesion area.
Using intelligent gastroenteroscopy operation assistance system, including data acquisition module, re-examination navigation module and re-examination image analysis module, the patient's digestive tract lesion location data and image data are analyzed through multi-objective genetic algorithm and neural network model to build the optimal examination path, reduce invasiveness, maximize regional coverage of lesions, and evaluate patient recovery.
Effectively reduce examination time, improve examination efficiency and patient comfort, provide scientific recovery evaluation results, help doctors formulate accurate treatment plans, improve treatment effects and health management levels.
Smart Images

Figure CN119028538A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of gastroenteroscopic operation assistance, and in particular to an intelligent gastroenteroscopic operation assistance system and method. Background Art
[0002] With the continuous advancement of medical technology, gastroenteroscopy, as a common diagnostic and treatment method, has higher and higher requirements on the accuracy and safety of its operation. Intelligent assistance systems can use advanced technology to improve the accuracy and efficiency of operators and reduce operational risks.
[0003] For example, the GasTro-AI intelligent colonoscopy-assisted diagnosis system based on YOLOv7 with publication number CN117637119A includes a computer, a terminal and a communication server; the computer includes a data preprocessing system, a lesion detection module, an automatic identification module and a patient case entry system, the communication server includes a server-side wireless communication module, a central processing module, a data cache module and a database, and the terminal includes a wireless network communication module, an image description system, a user interface development and a voice interaction system. Doctors or patients interact with the system through the terminal to view cases, medical histories, and exchange information, and receive treatment according to the treatment plan recommended by the system. Through the interconnection between the three parts of the computer, the communication server and the terminal and the various modules, the functions of lesion detection, case entry, doctor-patient communication, etc. are realized, providing doctors and patients with a convenient and efficient diagnosis and treatment platform.
[0004] However, the traditional gastroenteroscopy operation manual planning of the inspection path is difficult and prone to misoperation. Manual operation path planning takes a long time, which reduces the inspection efficiency. Manually planned paths may lead to unnecessary invasion, increasing the patient's discomfort and pain. Due to the lack of optimized inspection paths, multiple repeated inspections and unnecessary invasive operations may occur. The manual inspection path may not cover all lesion areas, and there is a risk of missed diagnosis. Different doctors' experience and operation methods may result in some lesion areas not being fully inspected. Summary of the invention
[0005] In view of the deficiencies in the prior art, the present invention provides an intelligent gastroenteroscopy operation assistance system and method, which solves the problems of the above-mentioned background technology.
[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: an intelligent gastroenteroscopy operation assistance system, including the following modules: a data acquisition module, a re-examination navigation module, a re-examination image analysis module, and a recovery evaluation module; the data acquisition module acquires the patient's historical gastrointestinal data through the hospital database, and the patient's historical gastrointestinal data includes the patient's digestive tract lesion location data and the patient's digestive tract historical image data; the re-examination navigation module is used to identify and mark the lesion area in the digestive tract based on the patient's digestive tract lesion location data, construct a gastroenteroscopy examination path, and use a multi-objective genetic algorithm to minimize the invasiveness of the patient's damage and the most accurate examination of the lesion area. The system can analyze the patient's digestive tract and the path of the gastrointestinal tract to obtain the optimal path of the gastrointestinal tract; the re-examination image analysis module is used to collect the patient's digestive tract re-examination image data through the gastrointestinal tract based on the optimal path of the gastrointestinal tract, extract features of the patient's digestive tract historical image data and the patient's digestive tract re-examination image data through the neural network model, analyze the lesion change characteristics in the lesion area image, and obtain the lesion change index; the recovery evaluation module is used to compare the lesion change index with the set lesion change index threshold, obtain the gastrointestinal recovery evaluation index, evaluate the patient's gastrointestinal recovery degree, and generate the patient's gastrointestinal recovery information.
[0007] Furthermore, the patient's digestive tract lesion location data includes: lesion site data, lesion type data, and lesion size data; the patient's digestive tract historical image data includes: time series data and lesion image feature data.
[0008] Furthermore, based on the patient's gastrointestinal lesion location data, the lesion area in the digestive tract is identified and marked, and the specific process of constructing a gastrointestinal endoscopic examination path is as follows: the lesion area in the digestive tract is automatically marked by an image segmentation algorithm, and a preliminary mark of the lesion area in the digestive tract is constructed; the patient's historical gastrointestinal images are aligned and compared using image registration technology, and the mark of the lesion area is updated and verified; a neural network model is used for feature extraction and classification, and the marked lesion area is classified and confirmed; based on the marked lesion area, a graph theory algorithm generates a preliminary gastrointestinal endoscopic examination path.
[0009] Furthermore, the specific process of generating a preliminary gastroenteroscopy examination path through a graph theory algorithm is as follows: the location of the lesion area is extracted from the image segmentation and labeling results, each lesion area is marked as a node in a graph, and the edges between the nodes are marked as the path for the gastroenteroscopy to move from one lesion area to another; the entry point of the gastroenteroscopy is set as the starting node, and the Dijkstra algorithm and minimum spanning tree are used to construct all the shortest paths and the best coverage of the lesion area path, respectively.
[0010] Furthermore, a multi-objective genetic algorithm was used to analyze the minimum invasiveness of patient injuries and the maximum coverage of the examination lesion area, and the gastrointestinal endoscopic examination path was screened. The specific process for obtaining the optimal gastrointestinal endoscopic examination path is as follows: the path length, invasion depth and invasion time are set as the minimum invasiveness constraint function, and the lesion area coverage and image quality are set as the maximum coverage objective function; the inspection path is represented as a chromosome through encoding, and the above-mentioned constraint function and objective function are used as fitness functions, and a new path group is generated through selection, crossover and mutation operations; the optimal path group is selected for retention according to the fitness function, and the above-mentioned genetic operations are iterated until the convergence conditions of satisfying the minimum invasiveness constraint and maximizing the coverage of the lesion area are reached, and the optimized optimal inspection path is output.
[0011] Furthermore, the lesion change characteristics in the lesion area image are analyzed, and the specific process of obtaining the lesion change index is as follows: the aligned image is feature extracted through the pre-trained convolutional neural network model, and the model can automatically extract high-dimensional features, extract the features of the re-examination image and the historical image, and compare them to identify the area change characteristics, color feature change characteristics and texture change characteristics of the lesion area; calculate the area difference between the lesion area in the re-examination image and the historical image, quantify the degree of lesion change, analyze the color feature change of the lesion area, quantify the differences in hue, brightness, and saturation characteristics, use the texture analysis algorithm, extract the texture features of the lesion area, and quantify the degree of change of the texture pattern; the area change characteristics, color feature change characteristics and texture change characteristics of the lesion area are comprehensively processed to obtain the lesion change index.
[0012] Furthermore, based on the lesion change index, the specific process of comparing with the set threshold of the lesion change index to obtain the gastrointestinal recovery assessment index is as follows: the difference between the lesion change index and the set threshold of the lesion change index is calculated, and a comprehensive operation is performed with the set allowable deviation value of the lesion change index to obtain the gastrointestinal recovery assessment index.
[0013] An intelligent gastroenteroscopy operation assistance method comprises the following steps: S1. Acquire historical gastrointestinal data of a patient through a hospital database, wherein the historical gastrointestinal data of the patient includes the location data of the patient's digestive tract lesions and the historical image data of the patient's digestive tract; S2. Based on the location data of the patient's digestive tract lesions, identify and mark the lesion area in the digestive tract, construct a gastroenteroscopy examination path, analyze the minimum invasiveness of the patient's injury and the maximum coverage of the lesion area through a multi-objective genetic algorithm, screen the gastroenteroscopy examination path, and obtain the optimal gastroenteroscopy examination path; S3. Based on the optimal gastroenteroscopy examination path, collect the patient's digestive tract re-examination image data through a gastroenteroscopy, extract features of the patient's digestive tract historical image data and the patient's digestive tract re-examination image data through a neural network model, analyze the lesion change characteristics in the lesion area image, and obtain the lesion change index; S4. Based on the lesion change index, compare with the set lesion change index threshold, obtain the gastrointestinal recovery evaluation index, evaluate the patient's gastrointestinal recovery degree, and generate the patient's gastrointestinal recovery information.
[0014] The present invention has the following beneficial effects:
[0015] (1) This intelligent gastroenteroscopy operation assistance system analyzes, identifies and marks the lesion area in the digestive tract by setting up a data acquisition module, a re-examination navigation module, a re-examination image analysis module, and a recovery evaluation module, constructs a gastroenteroscopy examination path, and uses a multi-objective genetic algorithm to analyze the minimum invasiveness of patient damage and the maximum coverage of the examination lesion area to select the gastroenteroscopy examination path. This path planning can effectively reduce the examination time, improve the examination efficiency, and reduce the patient's discomfort. Achieve efficient and accurate examination path planning, improve examination efficiency and patient comfort.
[0016] (2) This intelligent gastroenteroscopy operation assistance method analyzes the lesion change characteristics in the lesion area image, obtains the lesion change index, evaluates the patient's gastrointestinal recovery degree, and generates the patient's gastrointestinal recovery information. The systematic recovery evaluation method can provide scientific and quantitative evaluation results to help doctors formulate more accurate treatment plans. It can achieve scientific evaluation and management of the patient's recovery status and improve the treatment effect and the patient's health management level.
[0017] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS
[0018] Figure 1 This is a flow chart of an intelligent gastroenteroscopy operation assistance system of the present invention.
[0019] Figure 2 The present invention is a flow chart of an intelligent gastroenteroscopy operation auxiliary method. DETAILED DESCRIPTION
[0020] The embodiment of the present application reduces the patient's discomfort and pain through an intelligent gastroenteroscopy operation assistance system and method, realizes systematic analysis and storage of data, and comprehensively improves the quality of gastroenteroscopy examination and recovery evaluation.
[0021] The overall idea of the problem in the embodiment of this application is as follows:
[0022] The patient's historical gastrointestinal data is obtained through the hospital database, and the patient's historical gastrointestinal data includes the patient's digestive tract lesion location data and the patient's digestive tract historical image data.
[0023] Based on the patient's digestive tract lesion location data, the lesion area in the digestive tract is identified and marked, and a gastrointestinal endoscopic examination path is constructed. The multi-objective genetic algorithm is used to analyze the minimum invasiveness of patient damage and the maximum coverage of the lesion area, screen the gastrointestinal endoscopic examination path, and obtain the optimal gastrointestinal endoscopic examination path.
[0024] Based on the optimal examination path of gastroenteroscopy, the patient's digestive tract re-examination image data is collected through gastroenteroscopy, and the features of the patient's digestive tract historical image data and the patient's digestive tract re-examination image data are extracted through a neural network model. The lesion change characteristics in the lesion area image are analyzed to obtain the lesion change index.
[0025] Based on the lesion change index, it is compared with the set threshold of the lesion change index to obtain the gastrointestinal recovery assessment index, evaluate the patient's gastrointestinal recovery degree, and generate the patient's gastrointestinal recovery information.
[0026] See also Figure 1 The embodiment of the present invention provides a technical solution: an intelligent gastroenteroscopy operation auxiliary system, comprising the following modules: a data acquisition module, a re-examination navigation module, a re-examination image analysis module, and a recovery evaluation module; the data acquisition module acquires the patient's historical gastrointestinal data through a hospital database, and the patient's historical gastrointestinal data includes the patient's digestive tract lesion location data and the patient's digestive tract historical image data; the re-examination navigation module is used to identify and mark the lesion area in the digestive tract based on the patient's digestive tract lesion location data, construct a gastroenteroscopy examination path, analyze the minimum invasiveness of the patient's injury and the maximum coverage of the examination lesion area through a multi-objective genetic algorithm, screen the gastroenteroscopy examination path, and obtain the optimal gastroenteroscopy examination path; the re-examination image analysis module is used to collect the patient's digestive tract re-examination image data through a gastroenteroscopy based on the optimal gastroenteroscopy examination path, extract features of the patient's digestive tract historical image data and the patient's digestive tract re-examination image data through a neural network model, analyze the lesion change characteristics in the lesion area image, and obtain the lesion change index; the recovery evaluation module is used to compare the lesion change index with the set lesion change index threshold, obtain the gastrointestinal recovery evaluation index, evaluate the patient's gastrointestinal recovery degree, and generate the patient's gastrointestinal recovery information.
[0027] In this embodiment, the data acquisition module is a module for extracting and collecting the patient's historical gastrointestinal data from the hospital database. The information it obtains includes the location data of the patient's digestive tract lesions and the historical image data of the digestive tract. These data provide comprehensive background information for the system to support subsequent analysis and judgment. The re-examination navigation module is responsible for identifying and marking the lesion area in the digestive tract according to the lesion location data of the patient's digestive tract, and constructing a gastrointestinal endoscopy examination path. Through a multi-objective genetic algorithm, the module can analyze and select the optimal inspection path to ensure maximum coverage of the lesion area while minimizing damage to the patient. The multi-objective genetic algorithm is used to simultaneously consider multiple goals in the inspection path planning, minimize invasiveness and maximize coverage of the lesion area. The re-examination image analysis module uses a neural network model to collect and analyze the patient's re-examination image data through a gastrointestinal endoscopy based on the optimal inspection path. Through feature extraction and image analysis, the module evaluates the changes in the lesion area and generates a lesion change index. Neural network models are widely used in fields such as image processing and pattern recognition. In this system, the neural network model is used to extract features from the patient's historical image data and re-examination image data to analyze the changes in the lesion area. The lesion change index is used to represent the changes in the lesion area at different time points. By comparing the features of the patient's historical image data and re-examination image data, the lesion change index is calculated to evaluate the progression of the disease. Recovery Assessment Module This module generates a gastrointestinal recovery assessment index based on the lesion change index and compares it with the preset lesion change index threshold. It evaluates the patient's recovery and generates detailed recovery information to provide an effective reference for doctors. Image segmentation algorithm An image processing technology used to divide an image into several regions or target objects. In this system, the image segmentation algorithm is used to automatically mark the lesion area in the digestive tract and help construct preliminary markings of the lesion area in the digestive tract. Image registration technology is used to align and compare images taken at different time points or with different devices. In this system, image registration technology is used to align the patient's historical digestive tract images and re-examination images, and update and verify the markings of the lesion area. .
[0028] Specifically, the patient's digestive tract lesion location data includes: lesion site data, lesion type data, and lesion size data; the patient's digestive tract historical image data includes: time series data and lesion image feature data.
[0029] In this embodiment, the lesion location data indicates the exact location where the lesion occurs in the digestive tract. The lesion type data describes the specific type of lesion, which helps doctors assess the severity of the disease and choose appropriate treatment methods. Lesion size data refers to the size and range of the lesion area. The lesion size data helps doctors understand the extent of the lesion and evaluate the changes in the lesion during follow-up examinations. The time series data records the timestamps of the digestive tract image data taken at different time points. These time series data enable doctors to track the progression of lesions and the effectiveness of treatment, and understand the changes in lesions at different time points. Lesion image feature data: refers to feature data extracted from digestive tract images, including but not limited to features such as texture, shape, color, and edges. These feature data help the system and doctors to more accurately analyze and classify lesion areas and evaluate the nature and development trend of lesions.
[0030] Specifically, based on the patient's digestive tract lesion location data, the lesion area in the digestive tract is identified and marked, and the specific process of constructing a gastrointestinal endoscopic examination path is as follows: the lesion area in the digestive tract is automatically marked by an image segmentation algorithm, and a preliminary mark of the lesion area in the digestive tract is constructed; the patient's historical digestive tract images are aligned and compared using image registration technology, and the marks of the lesion area are updated and verified; a neural network model is used for feature extraction and classification, and the marked lesion area is classified and confirmed; based on the marked lesion area, a graph theory algorithm generates a preliminary gastrointestinal endoscopic examination path.
[0031] In this embodiment, the image segmentation algorithm refers to the use of computer vision technology to divide the medical image in the digestive tract into different areas, automatically identify and mark the lesion area. In the preliminary marking stage, the algorithm generates preliminary lesion area marks, which need to be verified and updated in subsequent steps. Image registration technology is to align images taken at different time points or different viewing angles for comparison and analysis. Common methods include feature point-based, density-based, and transformation model-based registration technologies. Update and verify that by aligning the current image and the historical image, the lesion area can be more accurately located, and the preliminary mark can be updated and verified to ensure the accuracy of the mark. Use the neural network model for feature extraction and classification, and classify and confirm the marked lesion area: use the neural network model in deep learning to process the image and extract high-dimensional features in the image; extract the features of the lesion area, texture, shape, color, etc., and then classify and confirm what type of lesions these areas are; classify the marked lesion area through the training and reasoning of the neural network model, and confirm its lesion nature; use the path planning algorithm in graph theory to generate the inspection path. The graph theory algorithm regards each lesion area in the digestive tract as a node in the graph, and the connection between the nodes represents the movement route of the gastroenteroscope. Based on these nodes and connections, an inspection path covering all lesion areas is generated. This path is preliminary and may need further optimization to ensure minimal invasiveness and maximum coverage.
[0032] Specifically, the specific process of generating a preliminary gastroenteroscopy examination path through a graph theory algorithm is as follows: the location of the lesion area is extracted from the image segmentation and labeling results, each lesion area is marked as a node in a graph, and the edges between the nodes are marked as the path for the gastroenteroscope to move from one lesion area to another; the entry point of the gastroenteroscope is set as the starting node, and the Dijkstra algorithm and minimum spanning tree are used to construct all shortest paths and the best coverage of the lesion area path, respectively.
[0033] In this embodiment, an image segmentation algorithm is used to automatically identify and mark the lesion area in the digestive tract. These marking results include the position of each lesion area in the image; each identified lesion area is regarded as a node in the graph theory. Each node represents a specific lesion location. The edges between the nodes are marked as the path for the gastroenteroscope to move from one lesion area to another: the edge represents the possible path for the gastroenteroscope to move from one lesion area to another. The starting position of the gastroenteroscope entering the digestive tract is set as the starting node in the graph theory. This is the starting point for path planning. The Dijkstra algorithm is used to find the shortest path from the starting node to all other nodes. The algorithm selects the node with the shortest path in turn for expansion until all nodes are visited, and constructs the shortest path from the starting node to each lesion area. The minimum spanning tree finds the optimal path covering all nodes to minimize the total path length. Common algorithms include the Kruska l algorithm and the Prim algorithm, which gradually add the shortest edges to ensure that no loops are formed, and finally generate the optimal path covering all lesion areas. The gastroenteroscope examination path is effectively planned by the graph theory algorithm to ensure that the lesion area is fully covered while minimizing the moving distance and time of the gastroenteroscope in the digestive tract. This not only improves the efficiency of the examination, but also reduces damage to the patient. The Dijkstra algorithm is used to determine the shortest path from the starting point to each lesion area, which helps to quickly locate the lesion. The minimum spanning tree algorithm ensures that all lesion areas are covered while keeping the total path length as short as possible, thereby optimizing the overall efficiency of the examination.
[0034] Specifically, the multi-objective genetic algorithm is used to analyze the minimum invasiveness of patient injuries and the maximum coverage of the examination lesion area, and the gastrointestinal endoscopic examination path is screened. The specific process of obtaining the optimal gastrointestinal endoscopic examination path is as follows: the path length, invasion depth and invasion time are set as the minimum invasiveness constraint function, and the lesion area coverage and image quality are set as the maximum coverage objective function; the inspection path is represented as a chromosome through encoding, and the above-mentioned constraint function and objective function are used as fitness functions, and a new path group is generated through selection, crossover and mutation operations; the optimal path group is selected for retention according to the fitness function, and the above-mentioned genetic operations are iterated until the convergence conditions of satisfying the minimum invasiveness constraint and maximizing the coverage of the lesion area are reached, and the optimized optimal inspection path is output.
[0035] In this implementation scheme, when using a multi-objective genetic algorithm to optimize the path of gastrointestinal endoscopy, the calculation formulas involved mainly include constraint functions and objective functions such as path length, invasion depth, invasion time, lesion area coverage, and image quality. The following are the specific calculation formulas involved. Path length calculation formula Path length refers to the total distance that the gastrointestinal endoscopy moves in the digestive tract. The path is P = p1, p2, ..., p n ], where p iis the position point in the digestive tract, the path length L(P) can be expressed as: Where, L(P): the total length of path P, which is composed of position points [p1,p2,…,p n Composition, p i : The i-th position point in the path represents a specific position of the gastroenteroscopy in the digestive tract, d(p i ,p i+1 ) represents the position point p in the path i and p i+1 The distance between the two, the penetration depth is calculated as follows: D(P) = max{d(p1,p i )||p i ∈P}; where P represents the path of gastrointestinal endoscopy, d(p1,p i ): The path starting point p1 and the i-th position point p in the path i The distance between them, max represents the maximum distance in the acquisition path, that is, the deepest position of the gastroenteroscope into the digestive tract. The invasion time calculation formula is as follows: In the formula, v represents the average speed of the gastroenteroscope, and T(P) represents the time required for the gastroenteroscope to complete the entire inspection path. The calculation formula for the lesion area coverage is as follows: In the formula, P represents the path of gastrointestinal endoscopy, B represents the set of lesion areas, including all areas marked as lesions, P∩B represents the number of lesion areas covered by path P, |B| represents the total number of lesion areas, and C(P) represents the proportion of lesion areas covered by path P. The image quality calculation formula is as follows: Where n represents the total number of locations in path P, q(p i ) represents the i-th position point p in the path iThe image quality score of path P is obtained by calculating the image quality score of all positions on path P. Q(P) represents the average image quality of all positions on path P. The fitness function calculation formula is as follows: F(P) = α·L(P)+β·D(P)+γ·T(P)-δ·C(P)-∈·Q(P); where L(P) represents the total length of path P, D(P) represents the maximum intrusion depth of path P, T(P) represents the total inspection time of path P, C(P) represents the lesion area coverage of path P, Q(P) represents the average image quality of path P, α, β, γ, δ, ∈ represent the corresponding weight coefficients, which are used to adjust the weights of each part in the fitness function, and F(P) represents the fitness value of path P, which comprehensively considers the path length, intrusion depth, inspection time, lesion coverage and image quality. Selection, crossover and mutation operations: Selection means selecting chromosomes with higher fitness to enter the next generation according to the fitness function. Crossover means performing a crossover operation on the selected chromosomes to generate a new chromosome combination. Mutation means performing a mutation operation on some chromosomes to increase the diversity of the population and avoid local optimal solutions. Iteration means repeatedly performing selection, crossover and mutation operations to generate new path groups. In each generation, the optimal path group is retained according to the fitness function. The above genetic operations are iterated until the convergence conditions are met, that is, the minimum invasive constraint and the optimization goal of maximizing the coverage of the lesion area. When the algorithm reaches the preset optimization generation or convergence conditions, the optimized optimal inspection path is output.
[0036] Specifically, the lesion change characteristics in the lesion area image are analyzed to obtain the lesion change index. The specific process is as follows: the aligned image is feature extracted through a pre-trained convolutional neural network model. The model can automatically extract high-dimensional features, extract the features of the re-examination image and the historical image, and compare them to identify the area change characteristics, color feature change characteristics and texture change characteristics of the lesion area; calculate the area difference between the lesion area in the re-examination image and the historical image, quantify the degree of lesion change, analyze the color feature change of the lesion area, quantify the differences in hue, brightness and saturation characteristics, use a texture analysis algorithm to extract the texture features of the lesion area, and quantify the degree of change of the texture pattern; the area change characteristics, color feature change characteristics and texture change characteristics of the lesion area are comprehensively processed to obtain the lesion change index.
[0037] In this embodiment, the historical image and the re-examination image are aligned through image registration technology to ensure that the lesion area at the same position is aligned in the two images. Use a pre-trained convolutional neural network model to extract features from the aligned image. Compare the features of the re-examination image and the historical image to identify changes in the lesion area. Compare the area difference of the lesion area in the two images. Color feature changes: Analyze the color features of the lesion area, including changes in hue, brightness, and saturation. Use a texture analysis algorithm to extract the texture features of the lesion area and analyze its changes. Calculate the quantitative value of each characteristic change, as follows, the area change characteristic calculation formula is as follows: In the formula, A h Represents the area of the lesion region in the historical image, A r represents the area of the lesion region in the re-examination image, and ΔA represents the change in the area of the lesion region, which is expressed as the relative difference between the lesion area in the re-examination image and the historical image. The color feature change calculation formula is as follows: Hue change In the formula, B h represents the brightness of the lesion area in the historical image, B r Indicates the brightness of the lesion area in the re-examination image, and ΔB indicates the brightness change, which is the relative difference in brightness between the re-examination image and the historical image. Brightness change In the formula, H h Indicates the hue of the lesion area in the historical image, H r Indicates the hue of the lesion area in the re-examination image, and ΔH indicates the amount of hue change, which is expressed as the relative difference in hue between the re-examination image and the historical image. Saturation change In the formula, S h Indicates the saturation of the lesion area in the historical image, S r represents the saturation of the lesion area in the re-examination image, ΔS represents the change in saturation, and is expressed as the relative difference in saturation between the re-examination image and the historical image. The texture change characteristic calculation formula is Where, T h Represents the texture features of the lesion area in the historical image, T r It represents the texture features of the lesion area in the re-examination image, and ΔT represents the texture change, which is expressed as the relative difference between the texture features in the re-examination image and the historical image. The formula for calculating the lesion change index is as follows: In the formula, represents the lesion change index, ΔA represents the area change, ΔH represents the hue change, ΔB represents the brightness change, ΔS represents the saturation change, ΔT represents the texture change, α represents the weight coefficient of the area change, β1 represents the weight coefficient of the hue change, β2 represents the weight coefficient of the brightness change, β3 represents the weight coefficient of the saturation change, and γ represents the weight coefficient of the texture change.
[0038] Specifically, based on the lesion change index, it is compared with the set threshold of the lesion change index to obtain the gastrointestinal recovery assessment index. The specific process is as follows: the difference between the lesion change index and the set threshold of the lesion change index is calculated, and the difference is combined with the set allowable deviation value of the lesion change index to obtain the gastrointestinal recovery assessment index.
[0039] In this embodiment, the calculation formula of the gastrointestinal recovery assessment index is as follows: Where ωf represents the gastrointestinal recovery assessment index, represents the lesion change index, ftg represents the threshold of the lesion change index, and Δfb represents the allowable deviation value of the lesion change index. The difference between the lesion change index and the threshold is calculated. The purpose of this step is to measure the difference between the actual lesion change and the expected change. The threshold is a set standard used to evaluate the expected range of lesion changes, and the allowable deviation value is used to define the acceptable range between the actual change and the expected change. Within this range, it means that the actual lesion change is consistent with the expected change.
[0040] See also Figure 2 , an intelligent gastroenteroscopy operation assistance method, comprising the following steps: S1. Acquire the patient's historical gastrointestinal data through the hospital database, the patient's historical gastrointestinal data including the patient's digestive tract lesion location data and the patient's digestive tract historical image data; S2. Based on the patient's digestive tract lesion location data, identify and mark the lesion area in the digestive tract, construct a gastroenteroscopy examination path, analyze the minimum invasiveness of the patient's injury and the maximum coverage of the lesion area through a multi-objective genetic algorithm, screen the gastroenteroscopy examination path, and obtain the optimal gastroenteroscopy examination path; S3. Based on the optimal gastroenteroscopy examination path, collect the patient's digestive tract re-examination image data through a gastroenteroscopy, extract features from the patient's digestive tract historical image data and the patient's digestive tract re-examination image data through a neural network model, analyze the lesion change characteristics in the lesion area image, and obtain the lesion change index; S4. Based on the lesion change index, compare with the set lesion change index threshold, obtain the gastrointestinal recovery evaluation index, evaluate the patient's gastrointestinal recovery degree, and generate the patient's gastrointestinal recovery information.
[0041] In this embodiment,.
[0042] In summary, this application has at least the following effects:
[0043] An intelligent gastroenteroscopy operation assistance system and method, by setting a data acquisition module, a re-examination navigation module, a re-examination image analysis module, and a recovery evaluation module, analyzes, identifies, and marks the lesion area in the digestive tract, constructs a gastroenteroscopy examination path, and uses a multi-objective genetic algorithm to analyze the minimum invasiveness of patient damage and the maximum coverage of the lesion area to screen the gastroenteroscopy examination path. This path planning can effectively reduce the examination time, improve the examination efficiency, and reduce the patient's discomfort. Achieve efficient and accurate inspection path planning, improve inspection efficiency and patient comfort, analyze the lesion change characteristics in the lesion area image, obtain the lesion change index, evaluate the patient's gastrointestinal recovery degree, and generate the patient's gastrointestinal recovery information. The systematic recovery evaluation method can provide scientific and quantitative evaluation results to help doctors formulate more accurate treatment plans. Achieve scientific evaluation and management of patient recovery, improve treatment effects and patient health management level.
[0044] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0045] The present invention is described with reference to flowcharts and / or block diagrams of systems, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0046] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0047] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.
[0048] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.
[0049] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.
Claims
1. An intelligent gastroenteroscopy operation assistance system, characterized in that: It includes the following modules: data acquisition module, re-examination navigation module, re-examination image analysis module, and recovery assessment module; The data acquisition module acquires the patient's historical gastrointestinal data through the hospital database, and the patient's historical gastrointestinal data includes the patient's digestive tract lesion location data and the patient's digestive tract historical image data; The re-examination navigation module is used to identify and mark the lesion area in the digestive tract based on the patient's digestive tract lesion location data, construct a gastrointestinal endoscopy examination path, analyze the minimum invasiveness of the patient's damage and the maximum coverage of the examination lesion area through a multi-objective genetic algorithm, screen the gastrointestinal endoscopy examination path, and obtain the optimal gastrointestinal endoscopy examination path; The re-examination image analysis module is used to collect the patient's digestive tract re-examination image data through the gastroenteroscopy based on the optimal examination path of the gastroenteroscopy, extract features from the patient's digestive tract historical image data and the patient's digestive tract re-examination image data through the neural network model, analyze the lesion change characteristics in the lesion area image, and obtain the lesion change index; The recovery assessment module is used to compare the lesion change index with a set threshold of the lesion change index based on the lesion change index, obtain a gastrointestinal recovery assessment index, assess the patient's gastrointestinal recovery degree, and generate the patient's gastrointestinal recovery information.
2. The intelligent gastroenteroscopy operation assistance system according to claim 1, characterized in that: The patient's digestive tract lesion location data includes: lesion location data, lesion type data, and lesion size data; The patient's digestive tract historical image data includes: time series data and lesion image feature data.
3. The intelligent gastroenteroscopy operation auxiliary system according to claim 2, characterized in that: Based on the patient's digestive tract lesion location data, the specific process of identifying and marking the lesion area in the digestive tract and constructing the gastrointestinal endoscopy examination path is as follows: Automatically mark the lesion area in the digestive tract through image segmentation algorithm, and construct preliminary marking of the lesion area in the digestive tract; Use image registration technology to align and compare the patient's historical digestive tract images, and update and verify the marking of the lesion area; Use neural network models for feature extraction and classification to classify and confirm the marked lesion areas; Based on the marked lesion areas, a graph theory algorithm generates a preliminary gastrointestinal endoscopy path.
4. The intelligent gastroenteroscopy operation assistance system according to claim 3, characterized in that: The specific process of generating a preliminary gastrointestinal endoscopy path through a graph theory algorithm is as follows: Extract the location of the lesion area from the image segmentation and labeling results, mark each lesion area as a node in a graph, and mark the edges between the nodes as a path for the gastroenteroscope to move from one lesion area to another; The entry point of the gastroenteroscopy was set as the starting node, and the Dijkstra algorithm and minimum spanning tree were used to construct all the shortest paths and the best path covering the lesion area, respectively.
5. The intelligent gastroenteroscopy operation auxiliary system according to claim 4, characterized in that: The specific process of analyzing the minimum invasiveness of patient damage and the maximum coverage of the lesion area through a multi-objective genetic algorithm, screening the gastrointestinal endoscopy examination path, and obtaining the optimal gastrointestinal endoscopy examination path is as follows: The path length, invasion depth and invasion time are set as the minimum invasive constraint function, and the lesion area coverage and image quality are set as the maximum coverage objective function; The inspection path is represented as a chromosome by encoding, the constraint function and the objective function are used as the fitness function, and a new path population is generated through selection, crossover and mutation operations; The optimal path group is selected for retention according to the fitness function, and the above genetic operation is iterated until the convergence conditions of satisfying the minimum invasiveness constraint and maximizing the coverage of the lesion area are reached, and the optimized optimal inspection path is output.
6. The intelligent gastroenteroscopy operation auxiliary system according to claim 5, characterized in that: The specific process of analyzing the lesion change characteristics in the lesion area image and obtaining the lesion change index is as follows: Through the pre-trained convolutional neural network model, the aligned images are subjected to feature extraction, high-dimensional features are automatically extracted, and features of the re-examination images and historical images are extracted and compared to identify the area change characteristics, color feature change characteristics, and texture change characteristics of the lesion area; Calculate the area difference between the lesion area in the re-examination image and the historical image, quantify the degree of change of the lesion, analyze the color feature changes of the lesion area, quantify the differences in hue, brightness, and saturation features, use texture analysis algorithms to extract the texture features of the lesion area, and quantify the degree of change of the texture pattern; The area change characteristics, color feature change characteristics and texture change characteristics of the lesion area are comprehensively processed to obtain the lesion change index.
7. The intelligent gastroenteroscopy operation assistance system according to claim 6, characterized in that: Based on the lesion change index, the specific process of comparing it with the set threshold of the lesion change index to obtain the gastrointestinal recovery assessment index is as follows: The difference between the lesion change index and the set threshold of the lesion change index is calculated, and a comprehensive calculation is performed with the set allowable deviation value of the lesion change index to obtain the gastrointestinal recovery assessment index.
8. An intelligent gastroenteroscopic operation assistance method, applied to an intelligent gastroenteroscopic operation assistance system according to any one of claims 1 to 7, characterized in that: The following steps are involved: S1. Obtain the patient's historical gastrointestinal data through the hospital database, where the patient's historical gastrointestinal data includes the patient's digestive tract lesion location data and the patient's digestive tract historical image data; S2. Based on the patient's digestive tract lesion location data, identify and mark the lesion area in the digestive tract, construct a gastrointestinal endoscopic examination path, analyze the minimum invasiveness of the patient's damage and the maximum coverage of the lesion area through a multi-objective genetic algorithm, screen the gastrointestinal endoscopic examination path, and obtain the optimal gastrointestinal endoscopic examination path; S3. Based on the optimal examination path of gastroenteroscopy, the patient's digestive tract re-examination image data is collected through the gastroenteroscopy, and the feature extraction of the patient's digestive tract historical image data and the patient's digestive tract re-examination image data is performed through the neural network model, and the lesion change characteristics in the lesion area image are analyzed to obtain the lesion change index; S4. Based on the lesion change index, compare it with the set threshold of the lesion change index to obtain the gastrointestinal recovery assessment index, evaluate the patient's gastrointestinal recovery degree, and generate the patient's gastrointestinal recovery information.
Citation Information
Patent Citations
Application of DICOM image in automatic navigation of endoscope capsule robot
CN114903415A
Digestive endoscopy navigation method and system
CN117204949A
Medical image diagnosis, comparison and reading method
CN118262875A
System and method for endoscopic path planning
US20050107679A1
Processing device, endoscope device, and processing method
WO2024095864A1
Cited By
Image focus recognition system for assisting digestive system department doctor in sampling inspection
CN119850632A
Image lesion recognition system to assist gastroenterologists in sampling and examination
CN119850632B
