Intraluminal Image Abnormality Detection via Integrated Feature Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image processing methods for intraluminal images struggle to accurately detect abnormalities such as tissue changes or lesions within the body, as they rely on limited feature data integration and detection criteria, leading to potential erroneous or missed detections.
Innovation Solution
An image processing device with an abnormality candidate region detection unit, feature data calculation unit, and integrated feature data calculation unit that detects and integrates multiple types of feature data (color, shape, texture) from intraluminal images to identify abnormal regions, using techniques like BoF and Fisher Vector for precise abnormality detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If limited feature data integration is used, then processing speed is maintained, but detection accuracy deteriorates
Solution Approach 1:
The patent segments feature data into multiple types (color features, shape features, texture features) and processes each type separately through dedicated calculation units, then integrates them systematically. This segmentation allows comprehensive feature analysis while maintaining organized processing flow, resolving the contradiction between detection accuracy and processing complexity.
Solution Approach 2:
The patent introduces a pyramid image structure that divides images into rectangular regions of multiple sizes, creating a multi-scale dimensional framework. This dimensional expansion allows feature data to be extracted and integrated at different scales, improving abnormality detection accuracy by capturing both local and global characteristics without overwhelming processing complexity.
2Measurement precision
If comprehensive feature data integration is implemented, then detection accuracy improves, but processing time increases
Solution Approach 1:
The patent performs preliminary actions by pre-calculating and storing representative vectors for different feature types before actual abnormality detection. The pyramid image structure and multiple feature types are prepared in advance, allowing the system to quickly compare incoming images against pre-established patterns, thereby reducing real-time processing time while maintaining comprehensive feature analysis.
3Reliability
If multiple feature types are integrated, then abnormality identification accuracy improves, but system complexity increases
Solution Approach 1:
The patent segments the system into specialized units for different feature types (color, shape, texture), each handled by dedicated calculation units. This modular segmentation improves reliability by ensuring each feature type is processed by appropriate algorithms while keeping system complexity manageable through clear functional separation and organized integration.
Data Source
AI summary
An image processing device includes: an abnormality candidate region detection unit configured to detect, from an intraluminal image obtained by imaging a living body lumen, an abnormality candidate region in which a tissue characteristic of the living body or an in-vivo state satisfies a predetermined condition; a feature data calculation unit configured to calculate, from each of a plurality of regions inside the intraluminal image, a plurality of pieces of feature data including different kinds; an integrated feature data calculation unit configured to calculate integrated feature data by integrating the plurality of pieces of feature data based on information of the abnormality candidate region; and a detection unit configured to detect an abnormality from the intraluminal image by using the integrated feature data.