Cluster Contrast for Unsupervised Person Re-Identification (2021-03-22T00:00:00.000000Z)

TL;DR

Cluster Contrast employs a unique cluster representation to describe each cluster, resulting in a cluster-level memory dictionary, which can solve the problem of cluster inconsistency and be applicable to larger data sets.

Abstract

State-of-the-art unsupervised re-ID methods train the neural networks using a memory-based non-parametric softmax loss. Instance feature vectors stored in memory are assigned pseudo-labels by clustering and updated at instance level. However, the varying cluster sizes leads to inconsistency in the updating progress of each cluster. To solve this problem, we present Cluster Contrast which stores feature vectors and computes contrast loss at the cluster level. Our approach employs a unique cluster representation to describe each cluster, resulting in a cluster-level memory dictionary. In this way, the consistency of clustering can be effectively maintained throughout the pipline and the GPU memory consumption can be significantly reduced. Thus, our method can solve the problem of cluster inconsistency and be applicable to larger data sets. In addition, we adopt different clustering algorithms to demonstrate the robustness and generalization of our framework. The application of Cluster Contrast to a standard unsupervised re-ID pipeline achieves considerable improvements of 9.9%, 8.3%, 12.1% compared to state-of-the-art purely unsupervised re-ID methods and 5.5%, 4.8%, 4.4% mAP compared to the state-of-the-art unsupervised domain adaptation re-ID methods on the Market, Duke, and MSMT17 datasets. Code is available at https://github.com/alibaba/cluster-contrast.

Authors

P. Tan

4 papers

Siyu Zhu

3 papers

Zuozhuo Dai

4 papers

TL;DR

Abstract

Authors

References73 items

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

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

Towards Discriminative Representation Learning for Unsupervised Person Re-identification

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

An Empirical Study of Training Self-Supervised Vision Transformers

ICE: Inter-instance Contrastive Encoding for Unsupervised Person Re-identification

Intra-Inter Camera Similarity for Unsupervised Person Re-Identification

Camera-aware Proxies for Unsupervised Person Re-Identification

CycAs: Self-supervised Cycle Association for Learning Re-identifiable Descriptions

Unsupervised Learning of Visual Features by Contrasting Cluster Assignments

Self-paced Contrastive Learning with Hybrid Memory for Domain Adaptive Object Re-ID

Unsupervised Person Re-Identification via Multi-Label Classification

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

Unsupervised Person Re-Identification via Softened Similarity Learning

Improved Baselines with Momentum Contrastive Learning

A Simple Framework for Contrastive Learning of Visual Representations

Mutual Mean-Teaching: Pseudo Label Refinery for Unsupervised Domain Adaptation on Person Re-identification

Self-Supervised Learning of Pretext-Invariant Representations

Momentum Contrast for Unsupervised Visual Representation Learning

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

Beyond Human Parts: Dual Part-Aligned Representations for Person Re-Identification

Self-Critical Attention Learning for Person Re-Identification

Unsupervised Graph Association for Person Re-Identification

Discriminative Feature Learning With Consistent Attention Regularization for Person Re-Identification

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

Contrastive Multiview Coding

Learning Representations by Maximizing Mutual Information Across Views

Generalizable Person Re-Identification by Domain-Invariant Mapping Network

Data-Efficient Image Recognition with Contrastive Predictive Coding

Unsupervised Embedding Learning via Invariant and Spreading Instance Feature

Invariance Matters: Exemplar Memory for Domain Adaptive Person Re-Identification

Local Aggregation for Unsupervised Learning of Visual Embeddings

Bag of Tricks and a Strong Baseline for Deep Person Re-Identification

Unsupervised Tracklet Person Re-Identification

Densely Semantically Aligned Person Re-Identification

Dissecting Person Re-Identification From the Viewpoint of Viewpoint

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

Batch DropBlock Network for Person Re-Identification and Beyond

Batch Feature Erasing for Person Re-identification and Beyond

Mancs: A Multi-task Attentional Network with Curriculum Sampling for Person Re-Identification

Generalizing a Person Retrieval Model Hetero- and Homogeneously

Unsupervised Person Re-identification by Deep Learning Tracklet Association

Learning deep representations by mutual information estimation and maximization

Two at Once: Enhancing Learning and Generalization Capacities via IBN-Net

Representation Learning with Contrastive Predictive Coding

Multi-task Mid-level Feature Alignment Network for Unsupervised Cross-Dataset Person Re-Identification

Pose Transferrable Person Re-identification

Unsupervised Feature Learning via Non-parametric Instance Discrimination

Transferable Joint Attribute-Identity Deep Learning for Unsupervised Person Re-identification

Person Transfer GAN to Bridge Domain Gap for Person Re-identification

Image-Image Domain Adaptation with Preserved Self-Similarity and Domain-Dissimilarity for Person Re-identification

Fine-Tuning CNN Image Retrieval with No Human Annotation

Random Erasing Data Augmentation

Sampling Matters in Deep Embedding Learning

Unsupervised Person Re-identification

In Defense of the Triplet Loss for Person Re-Identification

Re-ranking Person Re-identification with k-Reciprocal Encoding

Improved Texture Networks: Maximizing Quality and Diversity in Feed-Forward Stylization and Texture Synthesis

Performance Measures and a Data Set for Multi-target, Multi-camera Tracking

Large-scale vehicle re-identification in urban surveillance videos

Joint Detection and Identification Feature Learning for Person Search

Deep Residual Learning for Image Recognition

Scalable Person Re-identification: A Benchmark

FaceNet: A unified embedding for face recognition and clustering

Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift

ImageNet: A large-scale hierarchical image database

Maps of random walks on complex networks reveal community structure

Dimensionality Reduction by Learning an Invariant Mapping

A Density-Based Algorithm for Discovering Clusters in Large Spatial Databases with Noise

Supplementary Material for “Unsupervised Person Re-identiﬁcation by Soft Multilabel Learning”

Some methods for classification and analysis of multivariate observations

Some methods for classi cation and analysis of multivariate observations

John riccitiello sets out to identify the engine of growth for unity technologies ( interview ) . VentureBeat . Interview with Dean Takahashi

Field of Study

Journal Information

Name