Approximate Surface Calculation for Abnormal Region Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image processing technologies for detecting abnormal regions in in-vivo images, such as those from endoscopes, face challenges in accurately identifying changes in pixel values within the complex and varied environments of the human body, leading to potential misclassification of normal and abnormal regions.
Innovation Solution
An image processing apparatus and method that calculates multiple approximate surfaces to approximate pixel values, selects the most appropriate surface based on pixel value relations, sets an approximate region, and detects abnormal regions by comparing pixel values to the selected surface, thereby enhancing the accuracy of abnormal region detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional image processing technologies are used to detect abnormal regions, then the detection process is simple, but the accuracy of identifying abnormal regions is insufficient
Solution Approach 1:
The image processing method segments the examination-target region into multiple local regions and calculates approximate surfaces for each local region separately. This segmentation allows the system to handle complex variations in pixel values by treating different areas independently, thereby improving detection accuracy without requiring an overly complex global model.
Solution Approach 2:
The patent introduces a new dimension by calculating approximate surfaces (3D representation) from 2D pixel values. This dimensional transformation enables the system to model complex pixel value variations more effectively, improving the accuracy of abnormal region detection while maintaining a manageable processing framework.
2Measurement precision
If multiple approximate surfaces are calculated to improve detection accuracy, then the accuracy of abnormal region detection is improved, but the processing complexity increases
Solution Approach 1:
By dividing the examination-target region into multiple local regions and calculating approximate surfaces for each segment separately, the system achieves high processing accuracy through localized analysis. This segmentation strategy reduces the computational burden compared to calculating a single global approximate surface for the entire image, thereby managing processing time more efficiently.
Solution Approach 2:
The system calculates approximate surfaces for multiple local regions rather than attempting to create a single comprehensive model. This partial action approach focuses computational resources on creating accurate local approximations, which collectively provide high overall accuracy without requiring excessive processing time for a complete global model.
3Ease of manufacture
If color information is used for clustering to identify abnormal regions, then the method is easy to implement, but the accuracy of detecting abnormal regions with subtle pixel value changes is insufficient
Solution Approach 1:
The patent transitions from static color-based clustering to a dynamic approach by calculating approximate surfaces that model the relationships between pixel values. This dynamic modeling adapts to local variations in the image, enabling the detection of subtle abnormalities that static color clustering might miss, while maintaining implementation feasibility through systematic processing steps.
Solution Approach 2:
The system changes the parameter representation from simple color values to approximate surface parameters that capture the relationships between multiple pixel values. This parameter transformation enables more sensitive detection of subtle abnormalities by modeling the complex variations in pixel values, while the method remains implementable through structured calculation procedures.
Data Source
AI summary
An image processing apparatus includes an approximate-surface calculator that calculates multiple approximate surfaces that each approximate the pixel value of a pixel included in an examination-target region of an image; an approximate-surface selector that selects at least one approximate surface from the approximate surfaces on the basis of the relation between the pixel value of the pixel in the examination-target region and the approximate surfaces; an approximate-region setting unit that sets an approximate region that is approximated by at least the selected one approximate surface; and an abnormal-region detector that detects an abnormal region on the basis of the pixel value of a pixel in the approximate region and the value corresponding to the coordinates of that pixel on at least one approximate surface.