Pose-Driven Deep Convolutional Model for Person Re-identification

By Chi Su, Jianing Li, Shiliang Zhang, Junliang Xing, Wen Gao, Qi Tian. IEEE International Conference on Computer Vision, 2017

839 citations recorded by the source corpus.

Feature extraction and matching are two crucial components in person Re-Identification (ReID). The large pose deformations and the complex view variations exhibited by the captured person images significantly increase the difficulty of learning and matching of the features from person images. To overcome these difficulties, in this work we propose a Pose-driven Deep Convolutional (PDC) model to learn improved feature extraction and matching models from end to end. Our deep architecture explicitly leverages the human part cues to alleviate the pose variations and learn robust feature representations from both the global image and different local parts. To match the features from global human body and local body parts, a pose driven feature weighting sub-network is further designed to learn adaptive feature fusions. Extensive experimental analyses and results on three popular datasets demonstrate significant performance improvements of our model over all published state-of-the-art methods.

Subfields: Biomedical Engineering, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/fd5cbf72a0cc82542b543207a9d1ea382ca7cbcf

How researchers see this paper

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

In Biomedical Engineering

1. Yi Yang
2. Liang Zheng
3. Qi Tian
4. Zhedong Zheng
5. Shiliang Zhang
6. Shaogang Gong
7. Xiatian Zhu
8. Tao A. Xiang

In Computer Vision and Pattern Recognition

1. Yi Yang
2. Luc Van Gool
3. Dacheng Tao
4. Xiaogang Wang
5. Ming-Hsuan Yang
6. Ling Shao
7. Bernt Schiele
8. Qi Tian

Citations in the corpus

• Person Transfer GAN to Bridge Domain Gap for Person Re-identification (2017, methodology): “The Rank-1 accuracy on CUHK03 [20] labeled dataset has been improved from 19.9% [20] to 88.5% [27].”
• Pedestrian Alignment Network for Large-scale Person Re-Identification (2017, methodology): “In person re-ID, there are several works using body patch matching [53], [54].”
• A Pose-Sensitive Embedding for Person Re-identification with Expanded Cross Neighborhood Re-ranking (2017, methodology): “This correspondence can be established by explicitly using full body pose information for alignment [37] or locally through matching corresponding detected body parts [47, 48].”
• Self-Referenced Deep Learning (2018, result): “This is in strong contrast to existing deep re-id methods [44,40,28,32] where specially designed network architectures with complex training process are required in order to achieve the reported results.”
• CrowdPose: Efficient Crowded Scenes Pose Estimation and a New Benchmark (2018, background): “It has attracted tremendous interest for its wide applications in activity understanding [23, 15], human re-identification [27], human parsing [31, 12] etc.”
• 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.”
• Horizontal Pyramid Matching for Person Re-identification (2018, background): “In the first type, prior knowledge like poses or body landmarks are estimated and extracted to localize the discriminative regions (Su et al. 2017; Zheng et al. 2017; Wei et al. 2017).”
• Beyond Part Models: Person Retrieval with Refined Part Pooling (2017, methodology): “The latest state of the art on re-ID benchmarks are achieved with part-informed deep features [39,31,41].”
• Pose-Normalized Image Generation for Person Re-identification (2017, methodology): “For example, [39, 57] first detect normalized part regions from a person image, and then fuse the features extracted from the original images and the part region images.”
• ABD-Net: Attentive but Diverse Person Re-Identification (2019, background): “Many part-based methods have achieved superior performance [2, 3, 26, 27, 28, 4, 5, 29, 6, 30, 31, 8, 32].”
• AANet: Attribute Attention Network for Person Re-Identifications (2019, methodology): “tion and recognition accuracy can be significantly improved [18, 22, 24].”
• Applications of Generative Adversarial Networks (GANs): An Updated Review (2019, methodology): “This model differs significantly from the previous models [103, 142] [144] in that they synthesize realistic whole-body images using the proposed PN-GAN, rather than only focusing on body parts for pose normalization.”
• Perceive Where to Focus: Learning Visibility-Aware Part-Level Features for Partial Person Re-Identification (2019, background): “Recent works [23, 12, 26, 29, 22, 28, 16] further advance the state of the art on holistic person re-ID, through learning part-level deep features.”
• Part-Aligned Bilinear Representations for Person Re-identification (2018, methodology): “As the appearance network, [14] used GoogLeNet and [15] used ResNet50.”