Hierarchical Edge Detection in Image Processing Apparatus
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques for specifying edges of objects in images are not designed to accurately distinguish object edges from unnecessary edges, leading to inaccurate edge specification.
Innovation Solution
A method involving a computer-readable storage medium and image processing apparatus that sets larger and smaller regions in an image, calculates feature amounts, and determines edge regions by comparing these values to generate accurate edge image data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional edge detection techniques are used, then edges can be detected in the image, but many unnecessary edges are detected along with the object edges, reducing accuracy
Solution Approach 1:
The image is divided into multiple blocks, and each block is further divided into sub-blocks. This hierarchical segmentation allows the system to analyze edge characteristics at different spatial scales, distinguishing between significant object edges and insignificant dispersed edges through comparative analysis across block levels.
Solution Approach 2:
Different evaluation criteria are applied to different blocks based on their local characteristics. The system calculates feature amounts (such as variance or gradient magnitude) for each block and compares them to determine which blocks contain significant edges. This local quality assessment allows selective identification of important edges while filtering out noise.
2Reliability
If all edges are detected without discrimination, then comprehensive edge information is obtained, but the object edges cannot be accurately specified among the detected edges
Solution Approach 1:
The system changes the parameter of spatial scale by analyzing edges at multiple block levels. By comparing feature amounts across different block hierarchies and applying thresholding based on relative magnitudes, the system transforms the edge detection problem from a binary decision to a multi-level comparative analysis, improving reliability in identifying significant edges.
Data Source
AI summary
A non-transitory computer readable storage medium storing computer readable instructions executable by a computer is provided. The computer readable instructions cause the computer to obtain subject image data composing a subject image, set larger regions and smaller regions in each of the larger regions in the subject image, calculate a first feature amount of each of the smaller regions with use of values of pixels in each of the smaller regions and a second feature amount of each of the larger regions with use of values of pixels in each of the larger regions, determine whether each of the smaller regions is an edge region including an edge based on a comparison between the first feature amount and the second feature amount, and generate edge image data indicating edges in the subject image with use of results of the determination whether each of the smaller regions is an edge region.