Style Normalization and Restitution for Generalizable Person Re-Identification

@article{Jin2020StyleNA,
  title={Style Normalization and Restitution for Generalizable Person Re-Identification},
  author={Xin Jin and Cuiling Lan and Wenjun Zeng and Zhibo Chen and Li Zhang},
  journal={2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  year={2020},
  pages={3140-3149}
}
  • Xin Jin, Cuiling Lan, +2 authors Li Zhang
  • Published 2020
  • Computer Science
  • 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
Existing fully-supervised person re-identification (ReID) methods usually suffer from poor generalization capability caused by domain gaps. The key to solving this problem lies in filtering out identity-irrelevant interference and learning domain-invariant person representations. In this paper, we aim to design a generalizable person ReID framework which trains a model on source domains yet is able to generalize/perform well on target domains. To achieve this goal, we propose a simple yet… Expand
Unsupervised Constrative Person Re-identification
TLDR
This work proposes a novel selective contrastive learning framework for unsupervised feature learning, which proposes to use multiple positives and adaptively sampled negatives for defining the contrastive loss, enabling to learn a feature embedding model with stronger identity discriminative representation. Expand
Style Normalization and Restitution for DomainGeneralization and Adaptation
TLDR
A novel Style Normalization and Restitution module (SNR) is designed to simultaneously ensure both high generalization and discrimination capability of the networks and enforce a dual causality loss constraint in the restitution step to encourage the better separation of task-relevant and task-irrelevant features. Expand
Adaptive Domain-Specific Normalization for Generalizable Person Re-Identification
TLDR
This work proposes a novel adaptive domain-specific normalization approach (AdsNorm) for generalizable person Re-ID, which describes unseen target domain as a combination of the known source ones, and explicitly learns domain- specific representation with target distribution to improve the model’s generalization by a meta-learning pipeline. Expand
Generalizable Person Re-identification with Relevance-aware Mixture of Experts
TLDR
This work proposes a novel method called the relevance-aware mixture of experts (RaMoE), using an effective voting-based mixture mechanism to dynamically leverage source domains’ diverse characteristics to improve the model’s generalization. Expand
Dual-Stream Reciprocal Disentanglement Learning for Domain Adaption Person Re-Identification
TLDR
A novel method named Dual-stream Reciprocal Disentanglement Learning (DRDL), which is quite efficient in learning domain-invariant features and free from image generation, which removes redundant information from id-related features. Expand
Unsupervised and Self-Adaptative Techniques for Cross-Domain Person Re-Identification
TLDR
A novel UDA-based ReID method that optimizes a simple loss function with only one hyper-parameter and takes advantage of triplets of samples created by a new offline strategy based on the diversity of cameras within a cluster is proposed. Expand
Towards Discriminative Representation Learning for Unsupervised Person Re-identification
TLDR
This work proposes a unified framework by incorporating three technical schemes to address the challening unsupervised cross-domain re-ID problem, and proposes a cluster-wise contrastive learning algorithm by iterative optimization of feature learning and clustering to learn noise-tolerant feature representations. Expand
Semi-Supervised Domain Generalizable Person Re-Identification
TLDR
This work aims to explore multiple labeled datasets to learn generalized domain-invariant representations for person re-id, and proposes a simple yet effective Semi-Supervised Knowledge Distillation (SSKD) framework that effectively exploits the weakly annotated data by assigning soft pseudo labels to YouTube-Human for improving the generalization ability of models. Expand
Unsupervised Person Re-identification via Simultaneous Clustering and Consistency Learning
TLDR
This work designs a pretext task for unsupervised re-identification by learning visual consistency from still images and temporal consistency during training process, such that the clustering network can separate the images into semantic clusters automatically. Expand
Person30K: A Dual-Meta Generalization Network for Person Re-Identification
TLDR
A domain generalization ReID method, dual-meta generalization network (DMG-Net), to exploit the merits of meta-learning in both the training procedure and the metric space learning and designs a “learning then generalization evaluation” metatraining procedure and a meta-discrimination loss to enhance model generalization and discrimination capabilities. Expand
...
1
2
3
4
5
...

References

SHOWING 1-10 OF 78 REFERENCES
Generalizable Person Re-Identification by Domain-Invariant Mapping Network
TLDR
DIMN is designed to learn a mapping between a person image and its identity classifier, i.e., it produces a classifier using a single shot and can be used to match an arbitrary number of identities in a target domain. Expand
Frustratingly Easy Person Re-Identification: Generalizing Person Re-ID in Practice
TLDR
This paper proposes a simple baseline for domain generalizable~(DG) person re-identification, to learn a \reid{} model from a set of source domains that is suitable for application to unseen datasets out-of-the-box, without any model updating. Expand
Generalizing a Person Retrieval Model Hetero- and Homogeneously
TLDR
A Hetero-Homogeneous Learning (HHL) method is introduced that enforces two properties simultaneously: camera invariance, learned via positive pairs formed by unlabeled target images and their camera style transferred counterparts and domain connectedness, implemented by homogeneous learning. Expand
Semantics-Aligned Representation Learning for Person Re-identification
TLDR
A framework that drives the reID network to learn semantics-aligned feature representation through delicate supervision designs is proposed and achieves the state-of-the-art performances on the benchmark datasets CUHK03, Market1501, MSMT17, and the partial person reID dataset Partial REID. Expand
Horizontal Pyramid Matching for Person Re-identification
TLDR
A simple yet effective Horizontal Pyramid Matching (HPM) approach to fully exploit various partial information of a given person, so that correct person candidates can be still identified even even some key parts are missing. Expand
Invariance Matters: Exemplar Memory for Domain Adaptive Person Re-Identification
TLDR
Results on three re-ID domains show that the domain adaptation accuracy outperforms the state of the art by a large margin and the three invariance properties and the proposed memory are indispensable towards an effective domain adaptation system. Expand
Bag of Tricks and a Strong Baseline for Deep Person Re-Identification
TLDR
A simple and efficient baseline for person re-identification with deep neural networks by combining effective training tricks together, which achieves 94.5% rank-1 and 85.9% mAP on Market1501 with only using global features. Expand
Re-ID done right: towards good practices for person re-identification
TLDR
A qualitative analysis of the trained representation indicates that, while compact, it is able to capture information from localized and discriminative regions, in a manner akin to an implicit attention mechanism. Expand
Self-Similarity Grouping: A Simple Unsupervised Cross Domain Adaptation Approach for Person Re-Identification
TLDR
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. Expand
Resource Aware Person Re-identification Across Multiple Resolutions
TLDR
This work presents a new person re-ID model that combines effective embeddings built on multiple convolutional network layers, trained with deep-supervision, and proposes two new formulations of the person reID problem under resource-constraints. Expand
...
1
2
3
4
5
...