Medical Image Processing Apparatus for Breast Lesion Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current medical image analysis techniques, particularly in breast ultrasonic imaging, face challenges in accurately distinguishing lesion candidate regions from normal tissue, leading to high erroneous detection rates due to complex and unclear image patterns, and require costly anatomical structure analysis for real-time computer-aided detection.
Innovation Solution
A medical image processing apparatus that detects lesion candidate regions and evaluates their validity by extracting normal tissue regions based on structural features, calculating a variation feature amount between lesion and normal tissue features, and using this information to assess the detection result's validity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If anatomical structure analysis is performed to reduce erroneous detection, then detection accuracy is improved, but processing cost increases and real-time detection becomes difficult
Solution Approach 1:
The patent extracts only the necessary features (lesion candidate region features and surrounding region features) rather than performing complete anatomical structure analysis. By extracting specific relevant features and comparing them, the system achieves accurate detection without the high processing cost of full anatomical analysis.
Solution Approach 2:
The patent segments the image into lesion candidate regions and surrounding regions, then performs feature extraction and comparison on these segmented parts. This segmentation approach allows focused analysis on relevant areas rather than processing the entire anatomical structure, reducing computational complexity while maintaining detection accuracy.
2Measurement precision
If anatomical structure analysis is performed to reduce erroneous detection, then detection accuracy is improved, but processing time increases
Solution Approach 1:
The patent extracts only the necessary features (lesion candidate region features and surrounding region features) rather than performing complete anatomical structure analysis. By extracting specific relevant features and comparing them, the system achieves accurate detection without the high processing cost of full anatomical analysis.
Solution Approach 2:
The patent performs partial analysis by focusing only on lesion candidate regions and their immediate surrounding areas, rather than analyzing the entire anatomical structure. This partial action approach maintains detection accuracy while significantly reducing processing time to enable real-time detection.
3Reliability
If background region difference is used to reduce erroneous detection, then detection reliability is improved, but applicability to breast ultrasonic images is limited
Solution Approach 1:
The patent applies local quality by using the surrounding region (local area around lesion candidate) as the reference background instead of using distant background regions. This local reference approach is specifically adapted to breast ultrasonic images where mammary glands are localized, making the method both reliable and applicable to this specific imaging modality.
Data Source
AI summary
In medical examination of breast cancer, a lesion computer-aided detection is performed in real time and with high accuracy, and a burden on a medical worker is reduced. A medical image processing apparatus that processes a medical image includes: a detection unit configured to detect a lesion candidate region; a validity evaluation unit configured to evaluate validity of the lesion candidate region by using a normal tissue region corresponding to the detected lesion candidate region; and a display unit configured to determine display content to a user by using an evaluation result.