MHSA-Net: Multi-Head Self-Attention Network for Occluded Person Re-Identification
This addresses person re-identification, particularly under occlusion, for surveillance and security applications, but it appears incremental as it builds on existing attention-based methods.
The paper tackles the problem of occluded person re-identification by proposing MHSA-Net, which uses multi-head self-attention and an attention competition mechanism to capture key local information and filter noise, achieving competitive performance on standard and occluded tasks.
This paper presents a novel person re-identification model, named Multi-Head Self-Attention Network (MHSA-Net), to prune unimportant information and capture key local information from person images. MHSA-Net contains two main novel components: Multi-Head Self-Attention Branch (MHSAB) and Attention Competition Mechanism (ACM). The MHSAB adaptively captures key local person information, and then produces effective diversity embeddings of an image for the person matching. The ACM further helps filter out attention noise and non-key information. Through extensive ablation studies, we verified that the Multi-Head Self-Attention Branch (MHSAB) and Attention Competition Mechanism (ACM) both contribute to the performance improvement of the MHSA-Net. Our MHSA-Net achieves competitive performance in the standard and occluded person Re-ID tasks.