Endoscopic Image Selection for Lesion Documentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
During endoscopic examinations, doctors face the challenge of selecting relevant images from numerous captures for reports and AI qualitative determination, which is time-consuming and increases AI processing load, especially when multiple images relate to the same lesion.
Innovation Solution
An information processing device comprising an endoscopic image acquisition, still image acquisition, identity determination, and selection means to identify and select a representative image that best represents a lesion from multiple images, thereby streamlining the image selection process for reports and AI determination.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple still images are captured of the same lesion, then the completeness of lesion documentation is improved, but the time required for image selection and AI processing increases
Solution Approach 1:
The system enables self-service by automatically selecting representative images without requiring doctor intervention. The image selection unit autonomously identifies and selects the most representative still image from multiple captures of the same lesion, allowing the system to serve itself in the image selection task rather than relying on manual review by physicians.
Solution Approach 2:
The system changes the parameter of image representation by selecting a single representative image that best captures the essential features of the lesion. Instead of processing all captured images, the system transforms the dataset by identifying parameters that define representativeness (such as image quality, lesion visibility, and capture timing) and selects the image that optimizes these parameters.
2Measurement precision
If multiple still images of the same lesion are processed by AI, then the thoroughness of qualitative determination is improved, but the AI processing load increases
Solution Approach 1:
The system extracts only the essential representative image from multiple captures, removing redundant images that provide duplicate information. By taking out just the one most representative image for AI processing, the system maintains thoroughness of examination while eliminating unnecessary processing of redundant images, thereby reducing overall AI processing load.
Solution Approach 2:
Instead of processing all captured images (excessive action), the system applies partial action by processing only the single most representative image. This partial processing approach is sufficient to achieve thorough qualitative determination of the lesion while significantly reducing the computational burden compared to processing the entire set of captured images.
3Reliability
If dozens of images are captured during endoscopic examination, then the completeness of examination coverage is improved, but the complexity of image management increases
Solution Approach 1:
The system merges multiple images of the same lesion into a single representative image for reporting purposes. By combining the information from multiple captures and distilling it into one selected image, the system maintains complete examination coverage while simplifying image management and reducing the number of images that need to be stored, organized, and reviewed.
Solution Approach 2:
The system segments the large set of captured images by grouping images of the same lesion together and selecting one representative from each group. This segmentation approach divides the complex task of managing dozens of images into manageable units (lesion groups with selected representatives), thereby reducing overall image management complexity while preserving examination completeness.
Data Source
AI summary
The endoscopic image acquisition means acquires an endoscopic image. The still image acquisition means acquires a plurality of still images obtained by imaging a lesion included in the endoscopic image. The identity determination means determines an identity of the lesion included in the plurality of still images. The selection means selects a representative image that best represents the lesion, from a plurality of still images determined to correspond to the same lesion, and displays the representative image on the display device.