Proceedings ArticleDOI
Salience-Guided Cascaded Suppression Network for Person Re-Identification
Xuesong Chen,Canmiao Fu,Yong Zhao,Feng Zheng,Jingkuan Song,Rongrong Ji,Yi Yang +6 more
- pp 3300-3310
Reads0
Chats0
TLDR
A novel Salience-guided Cascaded Suppression Network (SCSN) which enables the model to mine diverse salient features and integrate these features into the final representation by a cascaded manner and develops an efficient feature aggregation strategy that fully increases the network’s capacity for all potential salience features.Abstract:
Employing attention mechanisms to model both global and local features as a final pedestrian representation has become a trend for person re-identification (Re-ID) algorithms. A potential limitation of these methods is that they focus on the most salient features, but the re-identification of a person may rely on diverse clues masked by the most salient features in different situations, e.g., body, clothes or even shoes. To handle this limitation, we propose a novel Salience-guided Cascaded Suppression Network (SCSN) which enables the model to mine diverse salient features and integrate these features into the final representation by a cascaded manner. Our work makes the following contributions: (i) We observe that the previously learned salient features may hinder the network from learning other important information. To tackle this limitation, we introduce a cascaded suppression strategy, which enables the network to mine diverse potential useful features that be masked by the other salient features stage-by-stage and each stage integrates different feature embedding for the last discriminative pedestrian representation. (ii) We propose a Salient Feature Extraction (SFE) unit, which can suppress the salient features learned in the previous cascaded stage and then adaptively extracts other potential salient feature to obtain different clues of pedestrians. (iii) We develop an efficient feature aggregation strategy that fully increases the network’s capacity for all potential salience features. Finally, experimental results demonstrate that our proposed method outperforms the state-of-the-art methods on four large-scale datasets. Especially, our approach exceeds the current best method by over 7% on the CUHK03 dataset.read more
Citations
More filters
Proceedings ArticleDOI
Combined Depth Space based Architecture Search For Person Re-identification
TL;DR: Wang et al. as discussed by the authors proposed a novel search space called Combined Depth Space (CDS), based on which they search for an efficient network architecture, which they call CDNet, via a differentiable architecture search algorithm.
Proceedings ArticleDOI
HAT: Hierarchical Aggregation Transformers for Person Re-identification
TL;DR: Zhang et al. as mentioned in this paper proposed a hierarchical aggregation transformer (HAT) for image-based person Re-ID, which takes advantage of both CNNs and Transformers to extract discriminative representations in a global view for persons under nonoverlapped cameras.
Posted Content
TransReID: Transformer-based Object Re-Identification
TL;DR: TransReID as mentioned in this paper proposes a pure transformer-based object ReID framework, which first encodes an image as a sequence of patches and builds a transformerbased strong baseline with a few critical improvements, which achieves competitive results on several ReID benchmarks.
Posted Content
Meta Batch-Instance Normalization for Generalizable Person Re-Identification
TL;DR: This paper proposes a novel generalizable Re-ID framework, named Meta Batch-Instance Normalization (MetaBIN), to generalize normalization layers by simulating unsuccessful generalization scenarios beforehand in the meta-learning pipeline, and shows that the model outperforms the state-of-the-art methods on the large-scale domain generalization Re- ID benchmark and the cross-domain Re-IDs problem.
Proceedings ArticleDOI
NTIRE 2021 NonHomogeneous Dehazing Challenge Report
Codruta Orniana Ancuti,Cosmin Ancuti,Florin-Alexandru Vasluianu,Radu Timofte,Minghan Fu,Huan Liu,Yankun Yu,Jun Chen,Keyan Wang,Jerome Chang,Xiyao Wang,Jing Liu,Yi Xu,Xinjian Zhang,Minyi Zhao,Shuigeng Zhou,Tianyi Chen,Jiahui Fu,Wentao Jiang,Chen Gao,Si Liu,Yudong Wang,Jichang Guo,Chongyi Li,Qixin Yan,Sida Zheng,Syed Waqas Zamir,Aditya Arora,Akshay Dudhane,Salman Khan,Munawar Hayat,Fahad Shahbaz Khan,Ling Shao,Haichuan Zhang,Tiantong Guo,Vishal Monga,Wenjin Yang,Jin Lin,Xiaotong Luo,Guowen Huang,Shuxin Chen,Yanyun Qu,Kele Xu,Lehan Yang,Pengliang Sun,Xuetong Niu,Junjun Zheng,Xiaotong Ruan,Yunfeng Wang,Jiang Yang,Zhipeng Luo,Sai Wang,Zhenyu Xu,Xiaochun Cao,Jun Luo,Zhuoran Zheng,Wenqi Ren,Tao Wang,Yiqun Chen,Cong Leng,Chenghua Li,Jian Cheng,Chang-Sung Sung,Jun-Cheng Chen,Eunsung Jo,Jae-Young Sim,Geethu M M,Akhil K A,Sreeni K G,Jeena R S,Joseph Zacharias,Chippy M Manu,Zexi Huang,Baofeng Zhang,Yiwen Zhang,Jindong Li,Mianjie Chen,Quan Xiao,Qingchao Su,Lihua Han,Yanting Huang,Kalpesh Prajapati,Vishal Chudasama,Heena Patel,Anjali Sarvaiya,Kishor P. Upla,Kiran B. Raja,Raghavendra Ramachandra,Christoph Busch,Hongyuan Jing,Zilong Huang,Yiran Fu,Haoqiang Wu,Quanxing Zha,Zhiwei Zhu,Hejun Lv +95 more
TL;DR: The results of the NTIRE 2021 Challenge on Non-Homogeneous Dehazing as mentioned in this paper have been evaluated on a novel dataset that consists of additional 35 pairs of real haze free and non-homogeneous hazy images recorded outdoor.
References
More filters
Proceedings ArticleDOI
Deep Residual Learning for Image Recognition
TL;DR: In this article, the authors proposed a residual learning framework to ease the training of networks that are substantially deeper than those used previously, which won the 1st place on the ILSVRC 2015 classification task.
Posted Content
Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
TL;DR: Faster R-CNN as discussed by the authors proposes a Region Proposal Network (RPN) to generate high-quality region proposals, which are used by Fast R-NN for detection.
Proceedings ArticleDOI
Rethinking the Inception Architecture for Computer Vision
TL;DR: In this article, the authors explore ways to scale up networks in ways that aim at utilizing the added computation as efficiently as possible by suitably factorized convolutions and aggressive regularization.
Posted Content
Rethinking the Inception Architecture for Computer Vision
TL;DR: This work is exploring ways to scale up networks in ways that aim at utilizing the added computation as efficiently as possible by suitably factorized convolutions and aggressive regularization.
Journal ArticleDOI
Squeeze-and-Excitation Networks
TL;DR: This work proposes a novel architectural unit, which is term the "Squeeze-and-Excitation" (SE) block, that adaptively recalibrates channel-wise feature responses by explicitly modelling interdependencies between channels and finds that SE blocks produce significant performance improvements for existing state-of-the-art deep architectures at minimal additional computational cost.