Semantic-Aware Occlusion-Robust Network for Occluded Person Re-Identification

Zhang, Xiaokang; Yan, Yan; Xue, Jing-Hao; Yang, Hua; Wang, Hanzi

doi:10.1109/tcsvt.2020.3033165

Cited by 57 publications

(14 citation statements)

References 51 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For instance, the richer data augmentation methods and the bounded exponential distance loss can improve the model robustness by obtaining more robust features from body parts that are not occluded, and the disentangled non‐local‐enabled backbone can also achieve better feature representation of the non‐occluded regions. For r=5 and r=10, the BMM method achieves the most superior values, which are 77.2% and 81.7%, while the suboptimal results that are achieved by the SORN [5] method are only 73.7% and 79.0%. This indicates the impressive performance of the our metric for occluded ReID.…”

Section: Resultsmentioning

confidence: 99%

“…In [8], a discriminative feature extraction and robust alignment based occluded person ReID method was proposed. In [5], a semantic‐aware model was presented, which included a branch for local feature extraction, a branch for occlusion‐robust global feature extraction and a branch for determination of the regions that were not occluded. In [9], the authors presented an approach via learning global features and reweighting partial features.…”

Section: Related Workmentioning

confidence: 99%

“…This further greatly increases the difficulty of person ReID because of the loss of discriminative information and the interference from the occluded regions. In this case, performance of the holistic ReID methods degrades sharply when occlusion occurs [5,6]. Therefore, it is in urgent need to specifically design the metric for occluded person ReID.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Bi‐level deep mutual learning assisted multi‐task network for occluded person re‐identification

Wang

Zhou

2022

IET Image Processing

View full text Add to dashboard Cite

An occluded person re-identification (ReID) approach is presented by constructing a Bi-level deep Mutual learning assisted Multi-task network (BMM), where the holistic and occluded person ReID tasks are treated as two related but not identical tasks. This is inspired by the human perception characteristic that there exist both similarities and differences when human views a holistic image and the occluded one. Specifically, a multi-task network with two branches is designed, where the convolutional neural network based feature representation part shares the weights by two tasks for commonality extraction, while the following output layers have respective weights for difference representation. Furthermore, as the non-occluded regions convey discriminative information, a bi-level mutual learning strategy is proposed and applied mutually on two branches to obtain more effective information from the non-occluded regions in the occluded images for better identity recognition. This is achieved by both feature-level and output-level mutual loss functions. Extensive experiments prove the advantages of the BMM for person ReID.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Bi‐level deep mutual learning assisted multi‐task network for occluded person re‐identification

Wang

Zhou

2022

IET Image Processing

View full text Add to dashboard Cite

show abstract

“…Semantic image segmentation, which assigns a label from a set of predefined classes to each pixel in an image, is a fundamental technique to characterize the contextual relationship of semantic categories in street scenes [1]. It can be used as a pre-processing step to remove uninformative regions [2], [3] or combined with 3D scene geometry [4], [5]. In general, these tasks require not only highresolution input images to cover a wide field of view, but also fast inference speed for interaction or response.…”

Section: Acknowledgmentmentioning

confidence: 99%

Stage-Aware Feature Alignment Network for Real-Time Semantic Segmentation of Street Scenes

Weng

Yan

Chen

et al. 2022

IEEE Trans. Circuits Syst. Video Technol.

Self Cite

View full text Add to dashboard Cite

Real-time performance is a very important trait of semantic segmentation models aiming at applications in robotics and intelligent transportation systems. Most previous work in the field involves custom convolutional encoders trained from scratch, and decoders without lateral skip-connections. However, we argue that a better speedaccuracy trade-off is achieved with i) compact encoders designed for competitive ImageNet performance and ii) lightweight decoders with lateral skip-connections. Additionally, we propose a novel interleaved pyramidal fusion scheme which is able to further improve the results on large objects close to the camera. We provide a detailed analysis of prediction accuracy and processing time on Cityscapes and CamVid datasets for models based on ResNet-18 and MobileNetv2. Our Cityscapes test submis-ages at 39.9 Hz on a GTX1080Ti. To the best of our knowledge, this result outperforms all previous approaches aiming at real-time application. The source code is available at https://github.com/orsic/swiftnet.

show abstract

“…Another network was presented in [63]. They have combined the intrinsic relationship between the tasks of person re-id and semantic segmentation to alleviate occlusion.…”

Section: Cnn-based Approachesmentioning

confidence: 99%

Person Re-identification: A Retrospective on Domain Specific Open Challenges and Future Trends

Zahra¹,

Perwaiz²,

Shahzad³

et al. 2022

Preprint

View full text Add to dashboard Cite

Person re-identification (Re-ID) is one of the primary components of an automated visual surveillance system. It aims to automatically identify/search persons in a multi-camera network having non-overlapping field-of-views. Owing to its potential in various applications and research significance, a plethora of deep learning based re-Id approaches have been proposed in the recent years. However, there exist several vision related challenges, e.g., occlusion, pose scale & viewpoint variance, background clutter, person misalignment and cross-domain generalization across camera modalities, which makes the problem of re-Id still far from being solved. Majority of the proposed approaches directly or indirectly aim to solve one or multiple of these existing challenges. In this context, a comprehensive review of current re-ID approaches in solving theses challenges is needed to analyze and focus on particular aspects for further advancements. At present, such a focused review does not exist and henceforth in this paper, we have presented a systematic challenge-specific literature survey of 230+ papers between the years of 2015-21. For the first time a survey of this type have been presented where the person re-Id approaches are reviewed in such solution-oriented perspective. Moreover, we have presented several diversified prominent developing trends in the respective research domain which will provide a visionary perspective regarding ongoing person re-Id research and eventually help to develop practical real world solutions. Scope/Objective of the ReviewIn this paper, we have targeted the most popular challenges in person re-id to perform systematic challenge-wise review of the published approaches. In this context, we have collected papers from top conferences and journals for the years from 2015 to 2021. The progress in papers addressing each challenge and its influence on published results is

show abstract

Semantic-Aware Occlusion-Robust Network for Occluded Person Re-Identification

Cited by 57 publications

References 51 publications

Bi‐level deep mutual learning assisted multi‐task network for occluded person re‐identification

Bi‐level deep mutual learning assisted multi‐task network for occluded person re‐identification

Stage-Aware Feature Alignment Network for Real-Time Semantic Segmentation of Street Scenes

Person Re-identification: A Retrospective on Domain Specific Open Challenges and Future Trends

Contact Info

Product

Resources

About