Abstract:Transformer-based supervised pre-training achieves great performance in person re-identification (ReID). However, due to the domain gap between ImageNet and ReID datasets, it usually needs a larger pre-training dataset (e.g. ImageNet-21K) to boost the performance because of the strong data fitting ability of the transformer. To address this challenge, this work targets to mitigate the gap between the pre-training and ReID datasets from the perspective of data and model structure, respectively. We first investigate self-supervised learning (SSL) methods with Vision Transformer (ViT) pretrained on unlabelled person images (the LUPerson dataset), and empirically find it significantly surpasses ImageNet supervised pre-training models on ReID tasks. To further reduce the domain gap and accelerate the pre-training, the Catastrophic Forgetting Score (CFS) is proposed to evaluate the gap between pre-training and fine-tuning data. Based on CFS, a subset is selected via sampling relevant data close to the down-stream ReID data and filtering irrelevant data from the pre-training dataset. For the model structure, a ReID-specific module named IBN-based convolution stem (ICS) is proposed to bridge the domain gap by learning more invariant features. Extensive experiments have been conducted to fine-tune the pre-training models under supervised learning, unsupervised domain adaptation (UDA), and unsupervised learning (USL) settings. We successfully downscale the LUPerson dataset to 50% with no performance degradation. Finally, we achieve state-of-the-art performance on Market-1501 and MSMT17. For example, our ViT-S/16 achieves 91.3%/89.9%/89.6% mAP accuracy on Market1501 for supervised/UDA/USL ReID. Codes and models will be released to https://github.com/michuanhaohao/TransReID-SSL.

Unleashing Potential of Unsupervised Pre-Training with Intra-Identity Regularization for Person Re-Identification

Supplementary: Unleashing Potential of Unsupervised Pre-Training with Intra-Identity Regularization for Person Re-Identification

Unsupervised Pre-training for Person Re-identification

Loose–tight cluster regularization for unsupervised person re-identification

Hybrid Dynamic Contrast and Probability Distillation for Unsupervised Person Re-Id

Reliability modeling and contrastive learning for unsupervised person re-identification

Fully Unsupervised Person Re-identification viaSelective Contrastive Learning

Self-Supervised Consistency Based on Joint Learning for Unsupervised Person Re-identification

A New Robust Contrastive Learning for Unsupervised Person Re-Identification

Unsupervised Person Re-identification via Simultaneous Clustering and Consistency Learning

3C: Confidence-Guided Clustering and Contrastive Learning for Unsupervised Person Re-Identification

Self-Supervised Pre-Training for Transformer-Based Person Re-Identification

Inter-Camera Identity Discrimination for Unsupervised Person Re-Identification

Hybrid Partial-Constrained Learning with Orthogonality Regularization for Unsupervised Person Re-Identification

Unified pre-training with pseudo infrared images for visible-infrared person re-identification

DCSG: Data Complement Pseudo-Label Refinement and Self-Guided Pre-Training for Unsupervised Person Re-Identification

Adaptive Attention-Aware Network for Unsupervised Person Re-Identification

Self-Training With Progressive Representation Enhancement for Unsupervised Cross-Domain Person Re-Identification

Hybrid Contrastive Learning with Cluster Ensemble for Unsupervised Person Re-identification

Large-Scale Unsupervised Person Re-Identification with Contrastive Learning

Take More Positives: An Empirical Study of Contrastive Learing in Unsupervised Person Re-Identification