A Research on the Application of Face Recognition Algorithm Based on Convolutional Model and Transformer Model in Community Monitoring System
0
조회
0
다운로드
1
인용
0
참고문헌
키워드
초록
The promotion of intelligent security community construction has greatly enhanced the intelligence and safety of residential areas. In order to further establish a security-oriented community, this paper proposes the utilization of facial recognition based on community surveillance footage to identify suspicious individuals. To address the difficulties in capturing facial images caused by factors such as low pixel resolution and varying shooting angles in surveillance footage, the following optimization strategies are proposed in this paper : Firstly, a lightweight global search facial detection network is designed based on convolutional modules and Vision Transformer modules. The Vision Transformer module is introduced to enhance the global retrieval capability of the network. Secondly, the structure of the Vision Transformer module is optimized by adding pooling layers in the feature block extraction and segmentation stage to reduce the number of module parameters. The feature blocks are mapped and computed with the feature maps to improve the corresponding feature correlation. Thirdly, in the face alignment stage, an Anchor Free mechanism is adopted to generate elliptical face localization regions for more accurate fitting of faces and reducing interference from other background information in the final identity recognition stage. Finally, the similarity between faces is calculated using Euclidean space distance to determine corresponding personnel identities. Through relevant experiments and tests on the self-built facial identity dataset in this paper's residential surveillance system, the proposed facial detection network achieves an average improvement of 3.11% in detection accuracy compared to other detection networks, reaching 97.19%. In terms of facial identity recognition, the designed model achieves an average improvement of 3.43% with a recognition accuracy of 95.84.
Abstract
The promotion of intelligent security community construction has greatly enhanced the intelligence and safety of residential areas. In order to further establish a security-oriented community, this paper proposes the utilization of facial recognition based on community surveillance footage to identify suspicious individuals. To address the difficulties in capturing facial images caused by factors such as low pixel resolution and varying shooting angles in surveillance footage, the following optimization strategies are proposed in this paper : Firstly, a lightweight global search facial detection network is designed based on convolutional modules and Vision Transformer modules. The Vision Transformer module is introduced to enhance the global retrieval capability of the network. Secondly, the structure of the Vision Transformer module is optimized by adding pooling layers in the feature block extraction and segmentation stage to reduce the number of module parameters. The feature blocks are mapped and computed with the feature maps to improve the corresponding feature correlation. Thirdly, in the face alignment stage, an Anchor Free mechanism is adopted to generate elliptical face localization regions for more accurate fitting of faces and reducing interference from other background information in the final identity recognition stage. Finally, the similarity between faces is calculated using Euclidean space distance to determine corresponding personnel identities. Through relevant experiments and tests on the self-built facial identity dataset in this paper's residential surveillance system, the proposed facial detection network achieves an average improvement of 3.11% in detection accuracy compared to other detection networks, reaching 97.19%. In terms of facial identity recognition, the designed model achieves an average improvement of 3.43% with a recognition accuracy of 95.84.
참고문헌 (15)
- W. Xin, "Fall Detection Algorithm for Workshop Workers Based on Visible-to-Thermal Infrared Visual Monitoring", Control and Decision, pp.1~9 (2023)
- Z. Dexiang, "Multi-scale All-scene Object Detection Method Based on Attention Mechanism", Journal of Electronics and Information Technology, 44(09), pp.3249~3257 (2022)
- X. Guishu, "Transformer-based Face Recognition Method with Multi-level Feature Fusion", Journal of Sichuan University (Natural Science Edition), pp.1~8 (2023)
- J. Ruirui, "Improved Visual Transformer-based Face Recognition Method", Journal of Computer Engineering and Applications, 59(08), pp.117~126 (2023)
- L. Jian, "A Survey on Transformer-based Object Detection Algorithms", Journal of Computer Engineering and Applications, 59(10), pp.48~64 (2023)
- L. Qingge, "A Survey on the Development of Transformer in Computer Vision", Small and Micro Computer Systems, 44(04), pp.850~861 (2023)
- L. Xiang, "A Survey on Transformer Research in Computer Vision", Journal of Computer Engineering and Applications, 59(01), pp.1~14 (2023)
- M. Yao, "A Survey on CNN and Transformer Applications in Fine-Grained Image Recognition", Journal of Computer Engineering and Applications, 58(19), pp.53~63 (2022)
- T. Shuang, "Light-weight Face Recognition Method with Feature Masking", Journal of Inner Mongolia University (Natural Science Edition), 54(03), pp.272~280 (2023)
- L. Yunhong, "Low-Resolution Face Recognition Based on Super-Resolution Reconstruction and Public Feature Subspace", Journal of Northwest University (Natural Science Edition), 53(02), pp.241~247 (2023)
- Y. Jie, "Masked Face Recognition Method Based on Lightweight Network", Electronic Measurement Technology, 46(06), pp.159~165 (2023)
- F. Schroff, "FaceNet : A Unified Embedding for Face Recognition and Clustering", IEEE (2015)[DOI]
- H. Wang, "CosFace : Large Margin Cosine Loss for Deep Face Recognition", computer vision and pattern recognition, pp.5265~5274 (2018)
- J. Deng, "ArcFace : Additive Angular Margin Loss for Deep Face Recognition", DOI:10.48550, 2018
- I. Perov, "DeepFaceLab : A simple, flexible and extensible face swapping framework", DOI:10.48550, 2020