Abstract:Clustering-based methods, which alternate between the generation of pseudo labels and the optimization of the feature extraction network, play a dominant role in both unsupervised learning (USL) and unsupervised domain adaptive (UDA) person re-identification (Re-ID). To alleviate the adverse effect of noisy pseudo labels, the existing methods either abandon unreliable labels or refine the pseudo labels via mutual learning or label propagation. However, a great many erroneous labels are still accumulated because these methods mostly adopt traditional unsupervised clustering algorithms which rely on certain assumptions on data distribution and fail to capture the distribution of complex real-world data. In this paper, we propose the plug-and-play graph-based pseudo label correction network (GLC) to refine the pseudo labels in the manner of supervised clustering. GLC is trained to perceive the varying data distribution at each epoch of the self-training with the supervision of initial pseudo labels generated by any clustering method. It can learn to rectify the initial noisy labels by means of the relationship constraints between samples on the k Nearest Neighbor (kNN) graph and early-stop training strategy. Specifically, GLC learns to aggregate node features from neighbors and predict whether the nodes should be linked on the graph. Besides, GLC is optimized with 'early stop' before the noisy labels are severely memorized to prevent overfitting to noisy pseudo labels. Consequently, GLC improves the quality of pseudo labels though the supervision signals contain some noise, leading to better Re-ID performance. Extensive experiments in USL and UDA person Re-ID on Market-1501 and MSMT17 show that our method is widely compatible with various clustering-based methods and promotes the state-of-the-art performance consistently.

Center Based Pseudo-Labeling For Semi-Supervised Person Re-Identification

Unlabeled Samples Generated by GAN Improve the Person Re-identification Baseline in Vitro

Online Pseudo Label Generation by Hierarchical Cluster Dynamics for Adaptive Person Re-identification

Label-Removed Generative Adversarial Networks Incorporating with K-Means

Learning From Synthetic Images via Active Pseudo-Labeling

Plug-and-Play Pseudo Label Correction Network for Unsupervised Person Re-identification

Two Birds with One Stone: Iteratively Learn Facial Attributes with GANs.

Cross-Model Pseudo-Labeling for Semi-Supervised Action Recognition

Dual Knowledge Distillation on Multiview Pseudo Labels for Unsupervised Person Re-Identification

Generative-Discriminative Complementary Learning

Label-Guided Generative Adversarial Network for Realistic Image Synthesis

Confidence-aware Pseudo-label Learning for Weakly Supervised Visual Grounding

GAN-CL: Generative Adversarial Networks for Learning from Complementary Labels

Complementary Pseudo Labels for Unsupervised Domain Adaptation on Person Re-identification.

Learning from Label Proportions with Generative Adversarial Networks

Deep Realistic Facial Editing via Label-restricted Mask Disentanglement

Adversarial Partial Multi-Label Learning

Pseudo labels purification for unsupervised person Re-IDentification

Cross-Domain Complementary Learning with Synthetic Data for Multi-Person Part Segmentation.

Lab2Pix: Label-Adaptive Generative Adversarial Network for Unsupervised Image Synthesis

Pseudo Label-Guided Model Inversion Attack Via Conditional Generative Adversarial Network.