Occluded Face Recognition Method Based on Multi-Scale Attention Mechanism
Authors:
- 20 September 2024
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 106-110
Abstract
In previous occluded face recognition models, the performance of partially occluded face recognition is significantly degraded due to the problems of limited recognizable features and high model complexity. In this article, We presented an occlusion method for faces based on a multi-scale attention mechanism by replacing the 3×3 convolution in the bottleneck block with a PSA attention module. The method cleans up the features damaged by masking by generating masks to conceal them, and then utilizes the multi-scale attention mechanism to effectively extract spatial information. Experiments show that the method is better than previous methods and effectively improves the accuracy.Keywords
This publication has 7 references indexed in Scilit:
- Generative adversarial networksCommunications of the ACM, 2020
- Occlusion Robust Face Recognition Based on Mask Learning With Pairwise Differential Siamese Network2019
- Robust LSTM-Autoencoders for Face De-Occlusion in the WildIEEE Transactions on Image Processing, 2017
- Occlusion robust face recognition based on mask learning2017
- Feature Pyramid Networks for Object Detection2017
- ImageNet classification with deep convolutional neural networksCommunications of the ACM, 2017
- Joint Face Detection and Alignment Using Multitask Cascaded Convolutional NetworksIEEE Signal Processing Letters, 2016