Face Verification Mode Switching for Occlusion and Low-Light Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing face verification technologies face challenges in accurately verifying faces under varying conditions, particularly with occlusions and in low light, and do not effectively balance usability and security in different verification scenarios.
Innovation Solution
A method and apparatus that utilize multiple neural network models trained on different data sets to extract feature information, with selective verification processes based on current modes, including normalization and occlusion handling, to enhance accuracy and adapt verification intensity for usability or security needs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple neural network models are used to extract feature information, then face verification accuracy is improved, but device complexity increases
Solution Approach 1:
The system divides the face verification task into multiple specialized neural network models, each trained on different data sets and optimized for specific conditions (e.g., occlusion handling, low light). This segmentation allows each model to excel at particular aspects of face verification while collectively providing comprehensive accuracy across diverse scenarios.
Solution Approach 2:
The system implements a universal verification framework that can dynamically select and combine multiple neural network models based on current verification modes and conditions. This multi-functional approach enables the system to adapt to various verification scenarios (usability-focused or security-focused modes) while maintaining a unified architecture that manages the complexity of multiple models.
2Reliability
If verification intensity is increased for security, then reliability is improved, but usability deteriorates
Solution Approach 1:
The system dynamically adjusts verification intensity and model selection based on the current verification mode. In security-focused mode, the system employs stricter verification criteria and may require multiple models to confirm identity, enhancing reliability. In usability-focused mode, the system uses more lenient criteria and fewer models, improving ease of operation. This dynamic adaptation allows the system to optimize the trade-off between security and usability based on contextual needs.
3Measurement precision
If normalization and occlusion handling are performed, then measurement precision is improved, but loss of time increases
Solution Approach 1:
The system performs normalization and occlusion handling as preliminary processing steps before feature extraction. By pre-processing the face image to correct for occlusions and standardize the input format, the system improves the accuracy of subsequent feature extraction while managing processing time through efficient preprocessing algorithms.
Solution Approach 2:
The system applies occlusion handling selectively to specific regions of the face image where occlusions are detected, rather than processing the entire image uniformly. This localized approach improves feature extraction accuracy in affected regions while minimizing the additional processing time required, as only problematic areas receive enhanced treatment.
Data Source
AI summary
A face verification method and apparatus is disclosed. The face verification method includes selecting a current verification mode, from among plural verification modes, to be implemented for the verifying of the face, determining one or more recognizers, from among plural recognizers, based on the selected current verification mode, extracting feature information from information of the face using at least one of the determined one or more recognizers, and indicating whether a verification is successful based on the extracted feature information.