Scalable Person Re-identification: A Benchmark

By Liang Zheng, Liyue Shen, Lu Tian, Shengjin Wang, Jingdong Wang, Qi Tian. 2015

4,560 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition

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

Learning Deep Disentangled Embeddings with the F-Statistic Loss (2018, methodology): “, 2014) and Market-1501 (Zheng et al., 2015)—following the methodology of Ustinova & Lempitsky (2016).”
Deep Mutual Learning (2017, methodology): “The Market-1501 [33] dataset is a widely used benchmark in the person re-identification (re-id) problem [5].”
In Defense of the Triplet Loss for Person Re-Identification (2017, methodology): “This means we directly optimize the network for the final task, which renders an additional metric learning step obsolete.”
Continual Representation Learning for Biometric Identification (2020, methodology): “Biometric identification [23,51], including face recognition [11, 37, 67] and person re-identification (re-id) [34, 66, 71], has achieved significant progress in the recent years due to the advances in modern learnable representations [8, 9, 11, 20, 34, 52, 60, 67] and emerging large datasets [17, 2”
Re-ranking Person Re-identification with k-Reciprocal Encoding (2017, methodology): “Most existing person reidentification methods mainly focus on feature representation [43, 13, 24, 50, 22] or metric learning [24, 18, 10, 33, 47].”
Random Erasing Data Augmentation (2017, neutral): “For person re-identification (re-ID), the Market-1501 dataset (Zheng et al. 2015) contains 12,936 images with 751 identities for training, 19,732 images with 750 identities and 3,368 query images for testing.”
Real-Time Multiple People Tracking with Deeply Learned Candidate Selection and Person Re-Identification (2018, background): “Market1501 [18], CUHK01 and CUHK03 [19], to enhance the generation ability for tracking.”
Omni-Scale Feature Learning for Person Re-Identification (2019, methodology): “Datasets and Settings We conduct experiments on six widely used person ReID datasets: Market1501 [68], CUHK03 [24], DukeMTMC-reID (Duke) [33, 70], MSMT17 [53], VIPeR [10] and GRID [29].”
Beyond Part Models: Person Retrieval with Refined Part Pooling (2017, methodology): “For Market-1501 and DukeMTMC-reID, we use the evaluation packages provided by [39] and [43], respectively.”
Pedestrian Alignment Network for Large-scale Person Re-Identification (2017, methodology): “We observe that the change of rank-1 accuracy and mAP are relatively Method rank-1 mAP BoW+kissme [Zheng et al., 2015] 25.13 12.17 LOMO+XQDA [Liao et al., 2015] 30.75 17.04 Gan [Zheng et al., 2017b] 67.68 47.13 OIM [Xiao et al., 2017] 68.1 - APR [Lin et al., 2017] 70.69 51.88 SVDNet [Sun et al.,…”
PifPaf: Composite Fields for Human Pose Estimation (2019, methodology): “To quantify the performance on the Market-1501 dataset, we created a simplified accuracy metric.”
End-to-End Comparative Attention Networks for Person Re-Identification (2016, methodology): “And for the Market-1501 dataset, mean average precision (mAP) as in [15] is also employed to evaluate the performance since there are on average 14.8 cross-camera ground truth matches for each query.”
AlignedReID: Surpassing Human-Level Performance in Person Re-Identification (2017, methodology): “Combining AlignedReID and mutual learning, our system outperforms state-of-the-art systems on Market1501, CUHK03, and CUHK-SYSU by a large margin.”