Masked Face Recognition With Layered Database Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing face recognition systems struggle to accurately identify individuals when they are wearing masks, leading to security vulnerabilities and health risks due to the need for mask removal, which is inconvenient and increases infection exposure.
Innovation Solution
A multi-layered approach for masked face recognition that includes database augmentation with simulated masked faces, real-time mask detection, and face restoration algorithms to reconstruct unmasked representations for accurate identification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional face recognition systems are used, then identification accuracy is maintained for unmasked faces, but recognition fails when users wear masks
Solution Approach 1:
The patent segments the face recognition process into multiple layers: (1) database augmentation layer that creates simulated masked face versions of existing unmasked face data, (2) real-time mask detection layer that identifies mask presence and type, and (3) masked face recognition layer that performs recognition adapted to the detected mask characteristics. This segmentation allows the system to handle both masked and unmasked faces effectively.
Solution Approach 2:
The patent performs preliminary actions by augmenting the face recognition database in advance with simulated masked face images generated from existing unmasked face data. This pre-processing creates a robust reference database that includes various mask types and positions, enabling the recognition system to accurately identify masked faces without requiring real-time mask removal or complex processing during actual recognition operations.
2Measurement precision
If masks are removed for authentication, then face recognition accuracy is improved, but health safety and user convenience deteriorate
Solution Approach 1:
The patent creates digital copies of unmasked face images by generating simulated masked face versions through image processing techniques. These synthetic masked face images are superimposed onto the original unmasked face data to create augmented database entries. This copying approach allows the system to learn from unmasked face data while operating effectively with masked faces, eliminating the need for physical mask removal.
3Adaptability or versatility
If mask detection and restoration algorithms are implemented, then masked face recognition capability is improved, but system complexity increases
Solution Approach 1:
The patent implements a dynamic multi-layered recognition system that adapts its processing approach based on real-time mask detection. The system dynamically selects and weights different recognition strategies: direct matching against simulated masked faces in the database, comparison of unmasked facial regions, and face restoration algorithms. This dynamic adaptation allows the system to optimize performance for different mask types and conditions while managing computational complexity through selective processing.
Solution Approach 2:
The patent incorporates feedback mechanisms where the results from different recognition layers are analyzed and selectively weighted to improve overall recognition accuracy. The system uses feedback from mask detection results to adjust the processing pipeline, and feedback from initial recognition attempts to refine face restoration. This iterative feedback approach enhances masked face recognition capability while maintaining system efficiency through intelligent resource allocation.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Embodiments of the present disclosure provide systems and methods for recognizing a masked face. According to the present disclosure, the disclosed systems and methods include features that provide augmentation of existing face recognition databases, real-time mask detection, and real-time masked face recognition. In embodiments, masked face recognition includes a multi-layered approach, which includes finding matching simulated masked faces in the database that match the masked face being analyzed, comparing the unmasked portion of the masked face to stored unmasked faces in a database to identify any matches, and executing face restoration algorithms in which the masked portion is reconstructed to generate an unmasked representation which may then be matched against unmasked faces in the database.