Face Verification Using Synthesized Images for Occlusion Handling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current face verification technologies face challenges in accurately verifying users when occluding objects such as glasses, sunglasses, hats, or masks are present, leading to false rejections due to differences in registration and verification feature extraction.
Innovation Solution
A processor-implemented method that generates a synthesized face image by replacing occluding object regions with reference image information, using a trained neural network to extract and compare face features, thereby reducing the impact of occlusions and improving verification accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If face verification is performed using traditional feature extraction methods, then the verification process is simple and fast, but the accuracy deteriorates when occluding objects are present
Solution Approach 1:
The patent segments the face image into multiple regions (occluded regions and non-occluded regions) and processes them differently. The feature extraction is divided into multiple stages: initial feature extraction from the original image, identification of occluded regions, generation of modified images by replacing occluded region features with synthetic features, and final feature extraction from multiple processed images. This segmentation approach improves verification accuracy under occlusion while managing system complexity through modular processing steps.
2Reliability
If occluding objects are detected and masking regions are applied, then verification accuracy under occlusion improves, but the processing time and computational complexity increase
Solution Approach 1:
The patent performs preliminary detection of occluding objects and identification of occluded regions before the main verification process. By detecting occlusions early and pre-processing the image to generate modified versions with synthetic feature replacements, the system prepares multiple candidate feature sets in advance. This preliminary action reduces the computational burden during the actual verification stage and minimizes processing time while maintaining high accuracy under occlusion.
3Adaptability or versatility
If multiple feature extraction methods are combined to handle occlusions, then the robustness against occlusion improves, but the device complexity increases
Solution Approach 1:
The patent merges multiple feature extraction approaches by combining features from the original face image with features from modified images where occluded regions have been replaced with synthetic features. The system integrates results from multiple processing paths (original image processing, occluded region detection, synthetic feature generation, and combined feature extraction) into a unified verification decision. This merging strategy enhances robustness against various occlusion types while managing complexity through integrated processing architecture.
Data Source
AI summary
A face verifying method and apparatus. The face verifying method includes detecting a face region from an input image, generating a synthesized face image by combining image information of the face region and reference image information, based on a determined masking region, extracting one or more face features from the face region and the synthesized face image, performing a verification operation with respect to the one or more face features and predetermined registration information, and indicating whether verification of the input image is successful based on a result of the performed verification operation.