Endoscopic Frame Tagging for Selective Image Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing endoscopic inspection methods struggle to efficiently extract important frame images from moving images due to insufficient image narrowing techniques and lack of user understanding of extraction criteria.
Innovation Solution
An endoscopic device with an operation receiving unit and processor that adds tags to frame images based on user operations and recognized features, allowing selective extraction of important frames using predefined tags.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If frame images are extracted based on multiple tag types (operation tags and feature image tags), then the accuracy and relevance of extracted images are improved, but the device complexity increases
Solution Approach 1:
The patent segments the frame image extraction process into two independent tagging systems: operation tags (marking user-operated frames) and feature image tags (marking frames with specific recognized features). This segmentation allows each tag type to be generated and processed independently, improving extraction accuracy through multi-criteria filtering while managing system complexity by modularizing the tagging functions.
2Reliability
If all frame images in a moving image are reviewed, then no important images are missed, but the time required for inspection increases
Solution Approach 1:
The system performs preliminary tagging of frame images during the recording process, marking frames with operation tags when user operations are detected and with feature image tags when specific features are recognized. This preliminary action prepares the data in advance, allowing rapid extraction of important frames during inspection without requiring review of all frames, thus reducing inspection time while maintaining completeness.
Solution Approach 2:
The patent introduces tags as intermediary markers that bridge the gap between raw frame images and important extracted images. These tags (operation tags and feature image tags) serve as mediators that automatically identify and flag significant frames, enabling inspectors to quickly locate important images without manually reviewing every frame, thereby reducing inspection time while ensuring no important images are missed.
3Measurement precision
If feature image recognition is performed on all frame images, then the relevance of extracted images is improved, but the processing time and computational resources increase
Solution Approach 1:
The patent applies local quality by performing feature image recognition selectively rather than uniformly on all frames. The system recognizes specific feature images (such as pyloric images, clip images, Vater's papilla images, and Bauhin valve images) only at relevant positions in the endoscopic sequence, allocating computational resources to frames where feature recognition is most valuable while reducing processing overhead for other frames.
Data Source
AI summary
A frame image extraction method including: during recording processing of a moving image including a plurality of frame images generated on a basis of an imaging signal output from an imaging element included in an insertion portion inserted into a test subject, among the plurality of frame images, adding, as a tag, information regarding an operation to a frame image corresponding to a timing when an operation receiving unit receives the operation, and adding, as a tag, information regarding a specific feature image to a frame image recognized as the specific feature image; and extracting a frame image from among the plurality of frame images included in the moving image on a basis of at least one type of tag among two or more types of tags selected from among a plurality of types of the tags.