Human Semantic Parsing for Person Re-identification

By Mahdi M. Kalayeh, Emrah Basaran, Muhittin Gokmen, Mustafa Ersel Kamasak, Mubarak Shah. 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2018

634 citations recorded by the source corpus.

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 state-of-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 state-of-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.

Subfields: Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/65da5a05fa6b49e8aa11b4acce965a0f439c1716

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

Citations in the corpus

ABD-Net: Attentive but Diverse Person Re-Identification (2019, neutral): “• SPReID [67] (2018 CVPR) 93.”
Perceive Where to Focus: Learning Visibility-Aware Part-Level Features for Partial Person Re-Identification (2019, methodology): “During matching stage, both PCB and SPReID adopt the common strategy of concatenating part features.”
Adaptive Transfer Network for Cross-Domain Person Re-Identification (2019, background): “Despite remarkable progress on person re-identification [26, 31, 11], it still remains a challenging task due to dramatic variances in imaging device, condition and environment among different surveillance cameras/camera networks.”
Hyperbolic Image Embeddings (2019, methodology): “A number of methods utilize a simple classification approach for training [3, 45, 17, 60], and Euclidean distance is used in test time.”
Deep Learning for Person Re-Identification: A Survey and Outlook (2020, methodology): “Human semantic parsing technique is also applied in [78] to capture the pixel-level discriminative cues and enhance the robustness against pose variations.”
Pose-Guided Feature Alignment for Occluded Person Re-Identification (2019, methodology): “All these approaches [15, 42, 24, 32] achieve better performance by assembling part-level features.”
Pose-Invariant Embedding for Deep Person Re-Identification (2017, background): “and testing sets are from the same dataset [34], [35], and for unsupervised domain adaptation [36], [37].”
Semantics-Enhanced Adversarial Nets for Text-to-Image Synthesis (2019, background): “In other computer vision tasks, such as person search [35], person Re-identification [17, 2], object tracking [11, 45] and Image Captioning [41, 22], visual attention mechanism been studied.”
A Strong Baseline and Batch Normalization Neck for Deep Person Re-Identification (2019, background): “Kalayeh et al. [23] proposed SPReID, which uses human semantic parsing to harness local visual cues for person ReID.”
Identity-Guided Human Semantic Parsing for Person Re-Identification (2020, methodology): “[18] propose to employ pretrained human parsing model to provide extra semantic.”
AANet: Attribute Attention Network for Person Re-Identifications (2019, methodology): “The three comparative networks are GPreID (Res-101) [1], SPReID (Res-152) [11] and the proposed AANet-152.”
Robust Person Re-Identification by Modelling Feature Uncertainty (2019, methodology): “Deep Person ReID Models Existing deep ReID models [2, 3, 7, 10, 18, 24, 27, 30, 33, 39] typically adopt an off-the-shelf CNN architecture for deep feature learning, with recent methods mostly using ResNet [6].”
Relation-Aware Global Attention for Person Re-Identification (2019, background): “In comparison with the attention based approaches [28, 25, 39, 19] which leverage human semantics ( e.g . foreground/background, human part segmentation) and those [22, 27, 32] which learn attention from input images themselves, our RGA-SC signiﬁcantly outperforms them.”