Refining Pseudo Labels with Clustering Consensus over Generations for Unsupervised Object Re-identification

  title={Refining Pseudo Labels with Clustering Consensus over Generations for Unsupervised Object Re-identification},
  author={Xiao Zhang and Yixiao Ge and Yu Qiao and Hongsheng Li},
  journal={2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  • Xiao Zhang, Yixiao Ge, Hongsheng Li
  • Published 1 June 2021
  • Computer Science
  • 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
Unsupervised object re-identification targets at learning discriminative representations for object retrieval without any annotations. Clustering-based methods [27], [46], [10] conduct training with the generated pseudo labels and currently dominate this research direction. However, they still suffer from the issue of pseudo label noise. To tackle the challenge, we propose to properly estimate pseudo label similarities between consecutive training generations with clustering consensus and… 

Figures and Tables from this paper

Part-based Pseudo Label Refinement for Unsupervised Person Re-identification

A novel Part-based Pseudo Label Refinement (PPLR) framework that reduces the label noise by employing the complementary relationship between global and part features and learns discriminative representations with rich local contexts.

Improving Unsupervised Domain Adaptive Re-Identification Via Source-Guided Selection of Pseudo-Labeling Hyperparameters

This work proposes new theoretical grounds on HP selection for clustering UDA re-ID as well as method of automatic and cyclic HP tuning for pseudo-labeling UDA clustering: HyPASS.

MGH: Metadata Guided Hypergraph Modeling for Unsupervised Person Re-identification

MGH is proposed, a novel unsupervised person ReID approach that uses meta information to construct a hypergraph for feature learning and label refinement, and takes advantage of label propagation on the hypergraph to effectively refine the ReID results.

Noise-Tolerant Learning with Silhouette Coefficient for Unsupervised Person Re-Identification

This work proposes a novel noise inhibition framework to estimate the confidence of each pseudo label and actively correct noisy labels and brings significant improvement and achieves the state-of-the-art performance on benchmark datasets.

A formal approach to good practices in Pseudo-Labeling for Unsupervised Domain Adaptive Re-Identification

A new theoretical view on Pseudo-Labeling UDA re-ID is proposed, formalized through a new general learning upper-bound on the UDA Re-Identification performance, and general good practices are directly deduced from the interpretation of the proposed theoretical framework in order to improve the target re- ID performance.

Offline-Online Associated Camera-Aware Proxies for Unsupervised Person Re-identification

This work proposes to split each single cluster into multiple proxies according to camera views, and designs two proxylevel contrastive learning losses that enable the combination of two losses to train a desirable Re-ID model.

Mind Your Clever Neighbours: Unsupervised Person Re-identification via Adaptive Clustering Relationship Modeling

A novel clustering relationship modeling framework for unsupervised person Re-ID where the relation between unlabeled images is explored based on a graph correlation learning (GCL) module and the refined features are used for clustering to generate high-quality pseudo-labels.

Multi-Centroid Representation Network for Domain Adaptive Person Re-ID

This paper presents a novel Multi-Centroid Memory (MCM) to adaptively capture different identity information within the cluster and proposes Second-Order Nearest Interpolation (SONI) to obtain abundant and informative negative samples.

Feature Diversity Learning with Sample Dropout for Unsupervised Domain Adaptive Person Re-identification

A brand-new method referred as to Feature Diversity Learning (FDL) under the classic mutual-teaching architecture, which can significantly improve the generalization ability of the feature representation on the target domain.



Hierarchical Clustering With Hard-Batch Triplet Loss for Person Re-Identification

The key idea of HCT is to make full use of the similarity among samples in the target dataset through hierarchical clustering, reduce the influence of hard examples through hard-batch triplet loss, so as to generate high quality pseudo labels and improve model performance.

Unsupervised Person Re-Identification via Multi-Label Classification

This paper formulates unsupervised person ReID as a multi-label classification task to progressively seek true labels by leveraging the updated ReID model for label prediction and proposes the memory-based multi- label classification loss (MMCL).

Unsupervised Person Re-Identification via Softened Similarity Learning

The iterative training mechanism is followed but clustering is discarded, since it incurs loss from hard quantization, yet its only product, image-level similarity, can be easily replaced by pairwise computation and a softened classification task.

A Bottom-Up Clustering Approach to Unsupervised Person Re-Identification

The experimental results demonstrate that the bottom-up clustering approach to jointly optimize a convolutional neural network and the relationship among the individual samples is not only superior to state-of-the-art unsupervised re-ID approaches, but also performs favorably than competing transfer learning and semi-supervised learning methods.

Unsupervised Person Re-Identification by Soft Multilabel Learning

The reference agent learning to represent each reference person by a reference agent in a joint embedding is introduced and the unified deep model outperforms the state-of-the-art unsupervised RE-ID methods by clear margins.

Self-Similarity Grouping: A Simple Unsupervised Cross Domain Adaptation Approach for Person Re-Identification

A Self-similarity Grouping (SSG) approach, which exploits the potential similarity of unlabeled samples to build multiple clusters from different views automatically, and introduces a clustering-guided semisupervised approach named SSG ++ to conduct the one-shot domain adaption in an open set setting.

Learning to Discover Novel Visual Categories via Deep Transfer Clustering

The problem of discovering novel object categories in an image collection is considered, and Deep Embedded Clustering is extended to a transfer learning setting, and the algorithm is improved by introducing a representation bottleneck, temporal ensembling, and consistency.

SLV: Spatial Likelihood Voting for Weakly Supervised Object Detection

A spatial likelihood voting (SLV) module is proposed to converge the proposal localizing process without any bounding box annotations, and an end-to-end training framework for multi-task learning is proposed.

AD-Cluster: Augmented Discriminative Clustering for Domain Adaptive Person Re-Identification

A novel augmented discriminative clustering (AD-Cluster) technique that estimates and augments person clusters in target domains and enforces the discrimination ability of re-ID models with the augmented clusters.

Exploring Bottom-Up and Top-Down Cues With Attentive Learning for Webly Supervised Object Detection

This work proposes a novel webly supervised object detection (WebSOD) method for novel classes which only requires the web images without further annotations, and combines bottom-up and top-down cues for novel class detection.