Exploring Spatial Significance via Hybrid Pyramidal Graph Network for Vehicle Re-Identification

Shen, Fei; Zhu, Jianqing; Zhu, Xiaobin; Xie, Yi; Huang, Jingchang

doi:10.1109/tits.2021.3086142

Cited by 92 publications

(23 citation statements)

References 87 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…And we run experiments on a NVIDIA V100 GPU with 16GB. Training configurations are summarized as follows [14]. (1)The input images are randomly sized to 224 × 224, 256 × 256, and 288 × 288.…”

Section: Implement Detailmentioning

confidence: 99%

See 1 more Smart Citation

A Competitive Method for Dog Nose-print Re-identification

Shen¹,

Wang²,

Wang³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Vision-based pattern identification (such as face, fingerprint, iris etc.) has been successfully applied in human biometrics for a long history. However, dog nose-print authentication is a challenging problem since the lack of a large amount of labeled data. For that, this paper presents our proposed methods for dog nose-print authentication (Re-ID) task in CVPR 2022 pet biometric challenge. First, considering the problem that each class only with few samples in the training set, we propose an automatic offline data augmentation strategy. Then, for the difference in sample styles between the training and test datasets, we employ joint cross-entropy, triplet and pair-wise circle losses function for network optimization. Finally, with multiple models ensembled adopted, our methods achieve 86.67% AUC on the test set. Codes are available at https: //github.com/muzishen/Pet-ReID-IMAG.

show abstract

Section: Implement Detailmentioning

confidence: 99%

“…In human biometrics, person/vehicle re-identification (Re-ID) [4,5,8,9,[11][12][13][14] methods based on deep learning have made a significant process in recent years. Pet biometric challenge 1 is a workshop in the ECCV2020 conference.…”

Section: Introductionmentioning

confidence: 99%

A Competitive Method for Dog Nose-print Re-identification

Shen¹,

Wang²,

Wang³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Because many off-the-shelf deep networks (e.g., ResNet [4], SeNet [5], and Res2Net [6]) can be applied to the global feature learning of vehicle re-identification via a few simple adjustments, i.e., inserting global average pooling and fully connected layers between off-the-shelf deep networks and loss functions. In the contrary, part feature learning methods [7,8,9,10,11,12] of vehicle re-identification are more complex, which require two steps, i.e., the part region locating step and the part feature extracting step. For both steps, there are a lot of methods [9,10,11,13] have been proposed.…”

Section: Introductionmentioning

confidence: 99%

Enhancing Part Features via Contrastive Attention Module for Vehicle Re-identification

Wei

et al. 2022

2022 IEEE International Conference on Image Processing (ICIP)

View full text Add to dashboard Cite

Vehicle re-identification's methods usually exploit the spatial uniform partition strategy via dividing deep feature maps into several parts. Then each of them is further independently processed by the multi-network branch to obtain refined part features. However, the cooperation among those part features is underestimated. This paper proposes a contrastive attention module (CAM) to assess one part feature's importance based on all parts. Practical cooperation among part features is derived by re-weighting the part feature. Furthermore, a flexible CAM network (CAMNet) compatible with contrastive attention module is proposed to enhance part features for vehicle re-identification. Extensive experiments show that the proposed CAMNet method outperforms many state-of-the-art vehicle re-identification approaches.

show abstract

“…Several recent works [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18] pay more attention to the local descriptors in the image and achieve the effect of enriching the representation ability of local descriptors by enhancing the spatial expression ability of convolution neural networks (CNN). The matching relationship [19], [20] is then obtained by simply calculating the distance between the local descriptors. However, this kind of method has two obvious shortcomings: 1) The above methods cannot provide structural information.…”

Section: Introductionmentioning

confidence: 99%

“…The proposed MASRG adopts a coarse-to-fine architecture as a whole, and introduces global feature information to guide the matching process of local descriptors, and contains two key modules, which are the structural relation graph module and the multi-scale attention module. Specifically, the structural relation graph module uses the global context information to introduce an adjacency matrix network with learnable parameters, re-weights it with local features, and re-models the adjacency relationship of the graph neural network aggregated by global and local descriptors, then the structural relationship between the descriptors is learned by a six-layer graph convolution neural network (GCN) [20], [24]. The structural relation graph module is composed of a joint relation module and a graph convolution network.…”

Section: Introductionmentioning

confidence: 99%

Multi-Scale Attention and Structural Relation Graph for Local Feature Matching

Nan

Ding

2022

IEEE Access

View full text Add to dashboard Cite

Building a dense correspondence between two images is a fundamental vision problem. Most existing methods use local features, but global features cannot be ignored. Local features are often not enough to disambiguate similar regions without global features. Computing relevant features between images requires structural relationship and the importance of local features. For that, We propose novel multi-scale attention and structural relation graph (MASRG) for local feature matching. The MASRG adopts an overall architecture that first builds coarse-level matches on a coarse feature map and then refines fine matches on a fine-level feature map. We propose a structural relation graph module and a multi-scale attention module. We introduce global context information into the overall architecture. Using global information to separately assist in learning the structural information between local descriptors, the features of different receptive fields, and the importance of modeling single local information, a limited number of possible matches can be obtained with high confidence. Finally, the matching relationship is predicted. In this way, the network significantly improves the matching reliability and localization accuracy. Our proposed method has 5.6%, 6.7%, and 6.3% performance increases over the baseline method(See I) under different conditions in the HPatches. Extensive experiments on three large-scale datasets (i.e., HPatches, InLoc, and Aachen Day-Night v1.1) demonstrate that our proposed MASRG method is superior to state-of-the-art local feature matching approaches.

show abstract

Exploring Spatial Significance via Hybrid Pyramidal Graph Network for Vehicle Re-Identification

Cited by 92 publications

References 87 publications

A Competitive Method for Dog Nose-print Re-identification

A Competitive Method for Dog Nose-print Re-identification

Enhancing Part Features via Contrastive Attention Module for Vehicle Re-identification

Multi-Scale Attention and Structural Relation Graph for Local Feature Matching

Contact Info

Product

Resources

About