Learning Feature Recovery Transformer for Occluded Person Re-Identification

Article Computer Science, Information Systems

Learning Disentangled Representation Implicitly Via Transformer for Occluded Person Re-Identification

Mengxi Jia et al.

Summary: This paper proposes a method for handling person re-IDentification under various occlusions. The DRL-Net network is designed to achieve alignment-free re-ID by reasoning global local features of occluded person images. It automatically disentangles the representation of undefined semantic components and measures image similarity. Additionally, a contrast feature learning technique (CFL) is designed to better separate occlusion features and discriminative ID features.

IEEE TRANSACTIONS ON MULTIMEDIA (2023)

Add to Collection

Article Computer Science, Artificial Intelligence

MHSA-Net: Multihead Self-Attention Network for Occluded Person Re-Identification

Hongchen Tan et al.

Summary: This article presents a novel person reidentification model, MHSA-Net, which prunes unimportant information and captures key local information to improve performance.

IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS (2023)

Add to Collection

Proceedings Paper Computer Science, Artificial Intelligence

Occlude Them All: Occlusion-Aware Attention Network for Occluded Person Re-ID

Peixian Chen et al.

Summary: The paper proposes an OcclusionAware Mask Network (OAMN) to address occluded person re-identification problems, which includes an attention-guided mask module requiring labeled occlusion data and a novel occlusion augmentation scheme. Experimental results demonstrate the superiority of OAMN over existing techniques on various person re-identification benchmark datasets.

2021 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2021) (2021)

Add to Collection

Article Biochemical Research Methods