Person30K: A Dual-Meta Generalization Network for Person Re-Identification

Jiao, Jile; Wang, Ce; Li, Jun; Lou, Yihang; Feng, Xuetao; Duan, Ling-Yu

doi:10.1109/cvpr46437.2021.00216

Cited by 51 publications

(17 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…When each point cloud sequence is treated as a domain, especially when the number of categories is unbalanced, it cannot give full play to its advantages. Alternatively, taking inspiration from [47], [48], we can view each scanned category as one domain to design class-agnostic tracking. Another solution is that we can use data generation to enhance the generalization in real-world scenarios.…”

Section: G Discussionmentioning

confidence: 99%

Towards Class-agnostic Tracking Using Feature Decorrelation in Point Clouds

Tian¹,

Li²,

Liu³

2022

Preprint

Self Cite

View full text Add to dashboard Cite

Single object tracking in point clouds has been attracting more and more attention owing to the presence of LiDAR sensors in 3D vision. However, existing methods based on deep neural networks mainly focus on training different models for different categories, which makes them unable to perform well in real-world applications when encountering classes unseen during the training phase. In this work, we turn our thoughts to a more challenging task in the LiDAR point clouds, namely class-agnostic tracking, where a general model is supposed to be learned to handle targets of both observed and unseen categories. In particular, we first investigate the class-agnostic performances of the state-of-the-art trackers via exposing the unseen categories to them during testing. It is found that when the distribution is shifted from observed to unseen classes, how to constrain the fused features between template and search region to maintain generalization is a key factor of class agnostic tracking. Therefore, we propose a feature decorrelation method to address this problem, which eliminates the spurious correlations of the fused features through a set of learned weights and further makes the search region consistent among foreground points and distinctive between foreground and background points. Experiments on the KITTI and NuScenes demonstrate that the proposed method can achieve considerable improvements by benchmarking against the advanced trackers P2B and BAT, especially when tracking unseen objects.

show abstract

Section: G Discussionmentioning

confidence: 99%

Towards Class-agnostic Tracking Using Feature Decorrelation in Point Clouds

Tian¹,

Li²,

Liu³

2022

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…The representative works are introduced as follows, respectively. 1) Inspired by meta-learning, MetaBIN (Choi et al 2021) stimulates and learns to handle the unsuccessful generalization situations, whereas DMG-Net (Bai et al 2021) is a dual-meta network to exploit the meta-learning more fully in both the training procedure and metric space learning.…”

Section: Related Workmentioning

confidence: 99%

“…Recently, generalizable re-id methods have drawn increasing attention, which can be roughly divided into three categories. The methods of the first category, based on metalearning, mimic the train-test splits to improve the ability of dealing with the stimulated generalization situations (Bai et al 2021;Choi et al 2021). The second methods aim to learn domain-invariant re-id features by the memory mechanism (Song et al 2019), hard example mining (Tamura and Murakami 2019) or adversarial learning (Chen et al 2021).…”

Section: Introductionmentioning

confidence: 99%

Generalizable Person Re-identification via Self-Supervised Batch Norm Test-Time Adaption

Han

Huang

et al. 2022

AAAI

View full text Add to dashboard Cite

In this paper, we investigate the generalization problem of person re-identification (re-id), whose major challenge is the distribution shift on an unseen domain. As an important tool of regularizing the distribution, batch normalization (BN) has been widely used in existing methods. However, they neglect that BN is severely biased to the training domain and inevitably suffers the performance drop if directly generalized without being updated. To tackle this issue, we propose Batch Norm Test-time Adaption (BNTA), a novel re-id framework that applies the self-supervised strategy to update BN parameters adaptively. Specifically, BNTA quickly explores the domain-aware information within unlabeled target data before inference, and accordingly modulates the feature distribution normalized by BN to adapt to the target domain. This is accomplished by two designed self-supervised auxiliary tasks, namely part positioning and part nearest neighbor matching, which help the model mine the domain-aware information with respect to the structure and identity of body parts, respectively. To demonstrate the effectiveness of our method, we conduct extensive experiments on three re-id datasets and confirm the superior performance to the state-of-the-art methods.

show abstract

“…Person re-identification (ReID) aims to associate individual pedestrians in a camera network with non-overlapping views. Recently, many methods (Sun et al 2018;Wang et al 2018;Dai et al 2021;He et al 2021b;Bai et al 2021;Zhao et al 2021;Wu, Zhu, and Gong 2022) based on deep learning have made significant progress in this field, including extracting more discriminative features and designing more suitable metrics. For example, OS-Net (Zhou et al 2019) and OSNet-AIN (Zhou et al 2021) designed a novel backbone that both consider the discriminative feature learning and the computational cost.…”

Section: Related Workmentioning

confidence: 99%

Uncertainty Modeling with Second-Order Transformer for Group Re-identification

Zhang

Lai

Feng

et al. 2022

AAAI

View full text Add to dashboard Cite

Group re-identification (G-ReID) focuses on associating the group images containing the same persons under different cameras. The key challenge of G-ReID is that all the cases of the intra-group member and layout variations are hard to exhaust. To this end, we propose a novel uncertainty modeling, which treats each image as a distribution depending on the current member and layout, then digs out potential group features by random samplings. Based on potential and original group features, uncertainty modeling can learn better decision boundaries, which is implemented by two modules, member variation module (MVM) and layout variation module (LVM). Furthermore, we propose a novel second-order transformer framework (SOT), which is inspired by the fact that the position modeling in the transformer is coped with the G-ReID task. SOT is composed of the intra-member module and inter-member module. Specifically, the intra-member module extracts the first-order token for each member, and then the inter-member module learns a second-order token as a group feature by the above first-order tokens, which can be regarded as the token of tokens. A large number of experiments have been conducted on three available datasets, including CSG, DukeGroup and RoadGroup. The experimental results show that the proposed SOT outperforms all previous state-of-the-art methods.

show abstract

Person30K: A Dual-Meta Generalization Network for Person Re-Identification

Cited by 51 publications

References 23 publications

Towards Class-agnostic Tracking Using Feature Decorrelation in Point Clouds

Towards Class-agnostic Tracking Using Feature Decorrelation in Point Clouds

Generalizable Person Re-identification via Self-Supervised Batch Norm Test-Time Adaption

Uncertainty Modeling with Second-Order Transformer for Group Re-identification

Contact Info

Product

Resources

About