Person Re-Identification Screening for Overlap-Occluded Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing person re-identification (ReID) systems face challenges in maintaining accuracy due to occlusion, illumination changes, and clothing differences between query and candidate person images, leading to incorrect identity certifications.
Innovation Solution
Implement a method to detect excessive overlap in person images and prevent identity authentication when overlap is significant, using a threshold-based system to ensure accurate identification by excluding partially occluded or heavily overlapping images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If person re-identification is performed using images with significant overlap between multiple persons, then the authentication process can proceed without interruption, but the accuracy of identity certification deteriorates due to occlusion and image quality degradation
Solution Approach 1:
The patent applies preliminary action by calculating the overlap degree between multiple persons in the image before performing person re-identification. The system determines in advance whether the overlap exceeds a threshold value, and only proceeds with authentication when the overlap is within acceptable limits, thereby preventing accuracy degradation from the start
Solution Approach 2:
The patent introduces an intermediary mechanism by using overlap degree calculation as a preliminary assessment step. This intermediary evaluation determines whether the image conditions are suitable for accurate authentication, acting as a gatekeeper between image acquisition and identity certification processes
2Productivity
If all detected person images are used for identity authentication, then the system operates efficiently with minimal processing, but erroneous certifications increase due to partially occluded or heavily overlapping images
Solution Approach 1:
The patent applies local quality by evaluating specific regions and conditions of person images individually. Instead of treating all images uniformly, the system assesses the overlap degree for each detected person and applies different processing decisions based on local image quality characteristics, excluding only those with excessive overlap
Solution Approach 2:
The patent uses parameter changes by introducing an overlap degree threshold parameter that controls the authentication process. By adjusting this parameter, the system can dynamically control which images are suitable for authentication, balancing between processing efficiency and authentication reliability
Data Source
AI summary
A determination method executed by a computer, the determination method including calculating an overlap degree between a first person and another person among a plurality of persons, in response to detecting that the plurality of persons are appearing in an image obtained by capturing a predetermined space; determining whether or not a partial image included in the image and indicating the first person is to be a target of person determination, according to the calculated overlap degree; determining which of candidates the first person is, by using the partial image, in response to determining that the partial image is to be the target of the person determination; and preventing the determining of which of the candidates the first person is, by using the partial image, in response to determining that the partial image is not to be the target of the person determination.