
The core message of this paper is to propose a novel unsupervised framework for visible-infrared person re-identification (VI-ReID) that addresses the challenges of noisy pseudo-labels and large modality gaps between visible and infrared images.


coremsg

unsupervised-visible-infrared-person-re-identification-via-pseudo-label-correction-and-modality-level-alignment


Unsupervised Visible-Infrared Person Re-Identification via Pseudo-Label Correction and Modality-Level Alignment


title_rewrite


An adaptive intra-class variation contrastive learning algorithm for unsupervised person re-identification that selects appropriate samples and outliers to dynamically update the memory dictionary based on the current learning capability of the model, leading to improved performance and faster convergence.


adaptive-intra-class-variation-contrastive-learning-for-efficient-and-robust-unsupervised-person-re-identification


Adaptive Intra-Class Variation Contrastive Learning for Efficient and Robust Unsupervised Person Re-Identification
