Face recognition methods, devices, systems, and storage media based on element detection.
By combining deep learning neural networks for encoding and decoding with predictive feature neural networks and IOU modules, the problem of low efficiency in face occlusion recognition in multi-face scenarios is solved, achieving efficient and accurate multi-face occlusion recognition and improving user experience.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CHINA GUANGFA BANK
- Filing Date
- 2023-03-29
- Publication Date
- 2026-05-26
AI Technical Summary
Existing technologies are ineffective in recognizing face occlusion types in situations with multiple faces, and are inefficient and provide a poor user experience.
By combining deep learning neural networks for encoding and decoding with predictive feature neural networks, along with an IOU module, a candidate box classification module, and a logical judgment module, facial elements are identified and classified through high-level and low-level feature fusion, and redundant information is filtered out, enabling simultaneous recognition of multiple occluded faces.
It supports simultaneous recognition of multiple faces with occlusion, improving recognition efficiency and accuracy, meeting the needs of real-time video review, and providing a better user experience.
Smart Images

Figure CN116189270B_ABST