Person Re-identification in the Wild

By Liang Zheng, Hengheng Zhang, Shaoyan Sun, Manmohan Chandraker, Yi Yang, Qi Tian. Computer Vision and Pattern Recognition, 2016

758 citations recorded by the source corpus.

This paper presents a novel large-scale dataset and comprehensive baselines for end-to-end pedestrian detection and person recognition in raw video frames. Our baselines address three issues: the performance of various combinations of detectors and recognizers, mechanisms for pedestrian detection to help improve overall re-identification (re-ID) accuracy and assessing the effectiveness of different detectors for re-ID. We make three distinct contributions. First, a new dataset, PRW, is introduced to evaluate Person Re-identification in the Wild, using videos acquired through six synchronized cameras. It contains 932 identities and 11,816 frames in which pedestrians are annotated with their bounding box positions and identities. Extensive benchmarking results are presented on this dataset. Second, we show that pedestrian detection aids re-ID through two simple yet effective improvements: a cascaded fine-tuning strategy that trains a detection model first and then the classification model, and a Confidence Weighted Similarity (CWS) metric that incorporates detection scores into similarity measurement. Third, we derive insights in evaluating detector performance for the particular scenario of accurate person re-ID.

Subfields: Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/8f0a3a6997b3f31719f56c499c5024414a5dd8bf

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

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

Citations in the corpus

• A Pose-Sensitive Embedding for Person Re-identification with Expanded Cross Neighborhood Re-ranking (2017, methodology): “The PRW dataset allows for an evaluation of the robustness of a re-id method to false positive or mis-aligned person detections.”
• Diverse Knowledge Distillation for End-to-End Person Search (2020, methodology): “Following the evaluation setting in (Xiao et al. 2017; Zheng et al. 2017; Chen et al. 2020a), we compare our proposed method with state-of-the-art methods on person search.”
• Multi-Task Learning with Low Rank Attribute Embedding for Multi-Camera Person Re-Identification (2018, background): “There are also approaches dedicated to person reidentification in large camera networks involving more than two cameras [35], [36], [37], [38], [39], [40], [41].”
• FairMOT: On the Fairness of Detection and Re-identification in Multiple Object Tracking (2020, background): “There has been remarkable progress on object detection (Ren et al. 2015; He et al. 2017; Zhou et al. 2019a; Redmon and Farhadi 2018) and re-ID (Zheng et al. 2017a; Chen et al. 2018a) respectively recently which in turn boosts the overall tracking accuracy.”
• Knowledge Distillation for End-to-End Person Search (2019, neutral): “The dispute reflects in literature arguing that detection and re-identification in person search should be addressed separately [7, 57, 68] or jointly [35, 39, 53, 55].”
• Person Re-identification: Past, Present and Future (2016, methodology): “CNN-based deep learning models have been popular since Krizhevsky et al. [14] won ILSVRC’12 by a large margin.”
• Re-ranking Person Re-identification with k-Reciprocal Encoding (2017, methodology): “PRW [54] is an end-to-end large-scale dataset.”
• Joint Object Detection and Multi-Object Tracking with Graph Neural Networks (2021, methodology): “Following TR-MOT [27], we use a mix of six pedestrian detection datasets for pre-training, including CalTech [70], MOT17 [65], CUHK-SYSU [32], ETH [71], PRW [72], and CityPersons [73].”
• Understanding and Controlling User Linkability in Decentralized Learning (2018, methodology): “All images in both datasets are publicly available on Flickr and have been used in various prior works [44, 51, 63, 72, 83].”
• SVDNet for Pedestrian Retrieval (2017, methodology): “In [32, 33], the classification CNN model is fine-tuned using either the video frames or image bounding boxes to learn a discriminative embedding for pedestrian retrieval.”
• Camera Style Adaptation for Person Re-identification (2017, methodology): “Another effective strategy is the classification model, which makes full use of the re-ID labels [43, 35, 29, 18, 36, 44, 41].”
• POI: Multiple Object Tracking with High Performance Detection and Appearance Feature (2016, background): “Such a dataset consists of multiple person re-id datasets, including PRW [18], Market-1501 [18], VIPeR [13] and CUHK03 [8].”
• Rethinking the Competition Between Detection and ReID in Multiobject Tracking (2020, methodology): “For a fair comparison, we use the same training datasets as the baseline tracker JDE [8], consisting of ETH [30], CityPerson [31], CalTech [32], MOT17 [33], CUDK-SYSU [34] and PRW [35].”
• Unlabeled Samples Generated by GAN Improve the Person Re-identification Baseline in Vitro (2017, methodology): “We adopt the CNN re-ID baseline used in [50, 51].”
• Pose-Invariant Embedding for Deep Person Re-Identification (2017, methodology): “Trained alone using a standard CNN architecture [4], [5], [12], we show that PoseBox yields very decent re-ID accuracy.”
• Deep Cosine Metric Learning for Person Re-identification (2018, background): “Nevertheless, several successful applications in face verification and person re-identification exist [24, 31, 37].”
• Learning Deep Context-Aware Features over Body and Latent Parts for Person Re-identification (2017, methodology): “[52] learn the IDE feature for the video-based person re-identification.”