Endoscopic AI Lesion Recognition With Standardized Staining
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current digestive endoscopy methods face challenges such as blurred vision due to mucous and foam, lengthy training periods for doctors, non-standardized staining solutions, and variations in concentration and methods, leading to high misdiagnosis rates and inefficiencies in diagnosing early digestive tract cancers.
Innovation Solution
A deep learning-based assistance system and device that integrates feature extraction, image classification, and early cancer recognition models, along with mucosal cleaning and staining technologies, to enhance diagnostic accuracy and efficiency by providing cleaner images and standardized staining.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If chromoendoscopy is used to improve lesion detection, then diagnostic accuracy is improved, but the process becomes more complex and time-consuming
Solution Approach 1:
The patent segments the complex chromoendoscopy process into distinct functional modules: a staining solution supply unit, a spraying unit with multiple nozzles, a rinsing unit, and a control unit. Each module performs a specific function, making the overall complex process more manageable and standardized
Solution Approach 2:
The patent standardizes the staining solution concentration parameter at 0.1%-0.5%, and controls spraying duration and rinsing parameters through the control unit. This parameter standardization reduces variability in the examination process while maintaining high diagnostic accuracy
2Measurement precision
If mucosal cleaning is performed to improve visualization, then diagnostic accuracy is improved, but examination time is increased
Solution Approach 1:
The patent applies preliminary action by pre-mixing the staining solution with foaming agents and surfactants that automatically clean the mucosal surface during the spraying process. This eliminates the need for separate cleaning steps, reducing examination time while maintaining visualization quality
Solution Approach 2:
The patent merges the mucosal cleaning function with the staining function by incorporating surfactants and foaming agents into the staining solution. The single spraying act simultaneously stains the mucosa and cleans away mucus and foam, eliminating separate cleaning steps
3Measurement precision
If standardized staining solutions are used to improve diagnostic consistency, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent designs a universal staining solution formulation (0.1%-0.5% concentration with surfactants and foaming agents) that works across different endoscope types and examination scenarios. The control unit provides universal control over spraying and rinsing parameters, reducing diagnostic variability without requiring multiple specialized systems
Solution Approach 2:
The patent introduces a control unit as an intermediary that manages the staining solution delivery and rinsing processes. This intermediary component standardizes the application parameters and timing, ensuring consistent diagnostic results while keeping the overall system integration straightforward
Data Source
Figure 1
Figure 2~4
Figure 5
AI summary
A deep learning-based examination and diagnosis assistance system and apparatus for early digestive tract cancer, the system comprising a feature extraction network, an image classification model, an endoscope classifier, and an early cancer recognition model; the feature extraction network is used for performing initial feature extraction on endoscope images on the basis of a neural network model; the image classification model is used for performing extraction on the initial features to acquire image classification features; the endoscope classifier is used for performing feature extraction on the initial features to acquire endoscope classification features and classify gastroscope or colonoscope images; the early cancer recognition model is used for splicing the initial features, the endoscope classification features, and the image classification features to acquire the probability of early cancer lesions in white light images, electronic dye images or chemical dye images of a corresponding site or acquire a flushing prompt or position recognition prompt for the corresponding site. The quality of AI-assisted diagnosis and the efficiency of digestive endoscopy diagnosis are thereby improved.