Circular Pattern Center Detection Using Weighted Hough Voting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting the center point of circular patterns in digital photographic images, such as pupil patterns, are prone to noise interference and are not effective for patterns with breakages, as they rely solely on edge data from Hough transform techniques.
Innovation Solution
The method involves voting on a Hough space defined by concentric annuluses using all pixel data from the image, integrating votes to identify the center point, and applying weighted voting based on chromatic color and luminance to suppress noise from non-relevant patterns, ensuring accurate detection of circular patterns with breakages.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If Hough transform techniques are used to detect circular patterns, then the ability to detect incomplete circular patterns and patterns with breakages is improved, but the detection becomes adversely affected by noise
Solution Approach 1:
The patent segments the circular pattern detection into multiple concentric annular regions (first annular region, second annular region, third annular region) with different weight coefficients. This segmentation allows different parts of the circular pattern to contribute differently to the Hough transform voting process, with the center region having higher weight to improve center point detection accuracy while edge regions have lower weight to reduce noise sensitivity.
Solution Approach 2:
The patent applies local quality by assigning different weight coefficients to different annular regions based on their distance from the center. The first annular region (center) receives a higher weight coefficient than the second and third annular regions (edges). This local differentiation ensures that the detection process focuses more on reliable center region data while reducing the influence of noisy edge data.
2Device complexity
If only edge pixel data is used for Hough transform voting, then the detection process is simpler, but the detection is apt to be adversely affected by noise
Solution Approach 1:
The patent segments the pixel data into multiple annular regions and applies different weight coefficients to each region. This segmentation transforms the simple edge detection approach into a multi-region weighted voting system that maintains computational feasibility while significantly improving noise resistance through differential weighting of reliable versus noisy regions.
3Reliability
If weighted voting based on chromatic color and luminance is applied, then noise suppression from non-relevant patterns is improved, but the processing complexity increases
Solution Approach 1:
The patent changes the voting parameters by introducing weight coefficients that are functions of chromatic color degree and luminance. Instead of uniform voting, each pixel's vote is weighted according to its color characteristics, with pixels having color degrees closer to the circular pattern's characteristic color receiving higher weights. This parameter transformation effectively suppresses noise from non-relevant patterns while maintaining processing efficiency through straightforward weight multiplication.
Data Source
AI summary
Voting of coordinates of each pixel in a photographic image onto a Hough space corresponding to annuluses is performed, the Hough space being defined by a coordinate system having an X coordinate axis for a circle center point, a Y coordinate axis for a circle center point, and an r coordinate axis for a radial direction. A vote value of each voting position having coordinates (X, Y, r) on the Hough space is acquired as a first integrated vote value from a calculation of a number of votes given to each voting position. A position represented by (X, Y) coordinate values of a voting position associated with the largest first integrated vote value is detected as the position of the center point of the circular pattern.