Person Re-identification Using Visual Attention

Rahimpour, Alireza; Liu, Liu; Taalimi, Ali; Song, Yang; Qi, Hairong

doi:10.48550/arxiv.1707.07336

Cited by 3 publications

(7 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Later, [20], [21], [19] simplify the attention scheme to integrate into CNN structures. Zhao et al [20] exploit the Spatial Transformer Network [35] as the hard attention model for searching discriminative parts given a pre-defined spatial constraint.…”

Section: Related Workmentioning

confidence: 99%

“…Thus, a simple human part-aligned representation is proposed for handling the body part misalignment problem. Rahimpour et al [21] introduce a gradient-based visual attention model, which learns to focus selectively on parts of the input image for which the networks' output is most sensitive to. Wang et al [19] present Residual Attention Network for image classification, built by stacking attention modules which generate attention-aware features, and bottom-up topdown feedforward structures which unfold the feedforward and feedback attention process into a single process.…”

Section: Related Workmentioning

confidence: 99%

“…We highlight the comparison with [19] because it performs best among the above-mentioned four methods (see Table IV -Table VI). Another reason is that it is easily compared with us, on the contrary, other methods [21], [18], [20] contain some additional techniques (such as RNN, STN [35]), which are not directly comparable to us. For direct and fair comparison between sharp attention and soft attention, we conduct experiments within the same setting: the same backbone network (i.e., ResNet-50), the same loss function (i.e., triplet loss), the same hyper-parameters and the same add-ons (i.e., CIL and CU).…”

Section: E Comparison With Soft Attention Modelsmentioning

confidence: 99%

“…More experimental comparison results with other soft attention methods on other datasets are detailed summarized in Table IV -Table VI. For [18], [21], [20], we just list the results reported in their papers. We find that the proposed method outperforms the soft attention models on all datasets, including CUHK03 labeled, CUHK03 detected and Market-1501.…”

Section: E Comparison With Soft Attention Modelsmentioning

confidence: 99%

“…Besides utilizing "very deep" network structures and employing different kinds of loss functions (e.g., Softmax [12], [13], triplet [14], [15] and Online Instance Matching [16]), a variety of solutions have been exploited intensively to enable more effective and efficient feature learning. Among these efforts is the attention mechanism focusing on the most discriminative parts of images in order to solve challenging recognition problems like person re-ID by distinguishing subtle finegrained visual structures from other irrelevant parts [17], [18], [19], [20], [21].…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Sharp Attention Network via Adaptive Sampling for Person Re-Identification

Shen

Jiang

et al. 2019

IEEE Trans. Circuits Syst. Video Technol.

View full text Add to dashboard Cite

In this paper, we present novel sharp attention networks by adaptively sampling feature maps from convolutional neural networks (CNNs) for person re-identification (re-ID) problem. Due to the introduction of sampling-based attention models, the proposed approach can adaptively generate sharper attention-aware feature masks. This greatly differs from the gating-based attention mechanism that relies soft gating functions to select the relevant features for person re-ID. In contrast, the proposed sampling-based attention mechanism allows us to effectively trim irrelevant features by enforcing the resultant feature masks to focus on the most discriminative features. It can produce sharper attentions that are more assertive in localizing subtle features relevant to re-identifying people across cameras. For this purpose, a differentiable Gumbel-Softmax sampler is employed to approximate the Bernoulli sampling to train the sharp attention networks. Extensive experimental evaluations demonstrate the superiority of this new sharp attention model for person re-ID over other related existing published state-of-theart works on three challenging benchmarks including CUHK03, Market-1501, and DukeMTMC-reID.Index Terms-person re-identification, sharp attention network, adaptive sampling, CNN.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: E Comparison With Soft Attention Modelsmentioning

confidence: 99%

Section: E Comparison With Soft Attention Modelsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Sharp Attention Network via Adaptive Sampling for Person Re-Identification

Shen

Jiang

et al. 2019

IEEE Trans. Circuits Syst. Video Technol.

View full text Add to dashboard Cite

show abstract

Human Semantic Parsing for Person Re-identification

Kalayeh

Başaran

Gokmen

et al. 2018

2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition

539

281

View full text Add to dashboard Cite

Person re-identification is a challenging task mainly due to factors such as background clutter, pose, illumination and camera point of view variations. These elements hinder the process of extracting robust and discriminative representations, hence preventing different identities from being successfully distinguished. To improve the representation learning, usually local features from human body parts are extracted. However, the common practice for such a process has been based on bounding box part detection. In this paper, we propose to adopt human semantic parsing which, due to its pixel-level accuracy and capability of modeling arbitrary contours, is naturally a better alternative. Our proposed SPReID integrates human semantic parsing in person re-identification and not only considerably outperforms its counter baseline, but achieves stateof-the-art performance. We also show that, by employing a simple yet effective training strategy, standard popular deep convolutional architectures such as Inception-V3 and ResNet-152, with no modification, while operating solely on full image, can dramatically outperform current stateof-the-art. Our proposed methods improve state-of-the-art person re-identification on: Market-1501 [48] by ∼17% in mAP and ∼6% in rank-1, CUHK03 [24] by ∼4% in rank-1 and DukeMTMC-reID [50] by ∼24% in mAP and ∼10% in rank-1.

show abstract

Face recognition using scattering convolutional network

Minaee

Abdolrashidi

Wang

2017

2017 IEEE Signal Processing in Medicine and Biology Symposium (SPMB)

View full text Add to dashboard Cite

Face recognition has been an active research area in the past few decades. In general, face recognition can be very challenging due to variations in viewpoint, illumination, facial expression, etc. Therefore it is essential to extract features which are invariant to some or all of these variations.Here a new image representation, called scattering transform/network, has been used to extract features from faces. The scattering transform is a kind of convolutional network which provides a powerful multi-layer representation for signals. After extraction of scattering features, PCA is applied to reduce the dimensionality of the data and then a multi-class support vector machine is used to perform recognition. The proposed algorithm has been tested on three face datasets and achieved a very high recognition rate.

show abstract

Person Re-identification Using Visual Attention

Cited by 3 publications

References 31 publications

Sharp Attention Network via Adaptive Sampling for Person Re-Identification

Sharp Attention Network via Adaptive Sampling for Person Re-Identification

Human Semantic Parsing for Person Re-identification

Face recognition using scattering convolutional network

Contact Info

Product

Resources

About