Keypoint Association Filtering via Statistical Similarity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image matching techniques face challenges in accurately distinguishing between correct and incorrect associations of keypoints due to noise effects and mismatching, which affects the reliability of image similarity assessment in computer vision applications.
Innovation Solution
The method employs statistical evaluation of similarity measures using logarithmic distance ratios (LDRs) and Mahalanobis distances, accounting for keypoint location uncertainties through covariance matrices, to differentiate between initial and global associations, thereby efficiently detecting incorrect associations between keypoints in images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If keypoint matching techniques are used to assess image similarity, then the capability to perform visual search and object recognition is enabled, but incorrect associations between keypoints occur due to noise effects and mismatching
Solution Approach 1:
The patent introduces an intermediary statistical evaluation mechanism that acts as a mediator between initial keypoint associations and final matching results. By computing similarity measures and comparing them against statistical thresholds derived from global associations, the system filters out incorrect associations while preserving correct ones, thus resolving the contradiction between enabling versatile image matching and ensuring association accuracy
Solution Approach 2:
The patent implements a feedback mechanism where the statistical properties of global associations (which predominantly contain incorrect associations) are used to establish thresholds that feed back into the evaluation of initial associations. This feedback loop enables the system to dynamically adjust its criteria for accepting or rejecting keypoint associations, improving reliability while maintaining matching capability
2Measurement precision
If statistical evaluation of similarity measures is performed to detect incorrect associations, then the accuracy of image matching is enhanced, but the computational complexity increases
Solution Approach 1:
The patent applies partial action by performing statistical evaluation selectively rather than exhaustively. Instead of evaluating all possible keypoint associations with full statistical rigor, the method computes similarity measures for initial associations and compares them against thresholds derived from a sampled set of global associations. This partial evaluation approach maintains detection accuracy while reducing computational complexity compared to exhaustive analysis
Data Source
AI summary
The invention relates to a method (100) for detecting incorrect associations between keypoints of a first image and keypoints of a second image, the method (100) comprising associating (101) every keypoint of the first image with every keypoint of the second image to obtain global associations, determining (103) similarity measures for pairs of initial associations, determining (105) similarity measures for pairs of global associations, determining (107) a first number of pairs of initial associations having a similarity measure within a predetermined range, determining (109) a second number of pairs of global associations having a similarity measure within the predetermined range, and comparing (111) the first number of pairs with the second number of pairs to detect the incorrect associations. The invention further relates to an apparatus for detecting incorrect associations between keypoints of a first image and keypoints of a second image.