Human Reidentification with Transferred Metric Learning

By Wei Li, Rui Zhao, Xiaogang Wang. Asian Conference on Computer Vision, 2012

679 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/44484d2866f222bbb9b6b0870890f9eea1ffb2d0

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. Xiaogang Wang
3. Liang Zheng
4. Luc Van Gool
5. Qi Tian
6. Dacheng Tao
7. Bernt Schiele
8. Tao A. Xiang

Citations in the corpus

• A Siamese Long Short-Term Memory Architecture for Human Re-identification (2016, neutral): “metric [28, 31, 33, 42, 57].”
• RGB-Infrared Cross-Modality Person Re-identification (2017, neutral): “) Datasets ID# images# cameras# RGB IR VIPER [7] 632 1,264 2 yes no iLIDS [54] 119 476 2 yes no CAVIAR [5] 72 610 2 yes no PRID2011 [11] 200 971 2 yes no CUHK01 [19] 972 1,942 2 yes no SYSU [8] 502 24,448 2 yes no CUHK03 [20] 1467 13,164 6 yes no Market [53] 1501 32,668 6 yes no MARS [52] 1261 1,191”
• Deep feature learning with relative distance comparison for person re-identification (2015, background): “Given a particular feature representation, a distance function is learned to construct a similarity measure [7, 8] with good similarity constraints .”
• Embedding Deep Metric for Person Re-identification: A Study Against Large Variations (2016, methodology): “Besides, to inspect the limitation of the data set CUHK01, we involve the recently released Market1501 [42] into the training.”
• Lifelong Person Re-Identification via Adaptive Knowledge Accumulation (2021, methodology): “Specifically, we first merge VIPeR [6], PRID [8], GRID [25], i-LIDS [41], CUHK01 [17], CUHK02 [16], SenseReID [50] in accordance with their original train/test splits as a new benchmark.”
• Attention-Aware Compositional Network for Person Re-identification (2018, methodology): “• Extensive experiments demonstrate that our approach achieve superior performance on several public datasets, including CUHK03 [25], CUHK01 [24], Market-1501 [55], CUHK03-NP [60], DukeMTMCreID [59] and SenseReID [49].”
• Locally Aligned Feature Transforms across Views (2013, background): “In [22], Li et al. proposed a transferred metric learning framework for learning specific metric for different query-candidate combinations.”
• Beyond frontal faces: Improving Person Recognition using multiple cues (2015, background): “Existing work is mainly focused on metric learning [27, 17, 22], mid-level feature learning [16, 39, 40, 10, 20, 26].”
• End-to-End Deep Learning for Person Search (2016, methodology): “1http://www.ee.cuhk.edu.hk/˜xgwang/PS/dataset. html Although numerous re-id datasets and algorithms have been proposed in recent years and the performance on these benchmarks have been improved substantially, there is still a big gap with practical applications.”
• Hierarchical Gaussian Descriptor for Person Re-identification (2016, methodology): “We compare the GOG descriptor with other descriptors using the four datasets: VIPeR [12], CUHK01 [20], GRID [26] and PRID450S [33], respectively contains 632, 971, 250 and 450 images of individuals captured in two disjoint camera views.”
• Cross Dataset Person Re-identification (2014, positive): “We apply almost the same structure and parameters of CUHK02 except the epoches.”
• Person Re-identification by Salience Matching (2013, methodology): “the VIPeR dataset [7], and the CUHK Campus dataset [12].”
• Robust Person Re-Identification by Modelling Feature Uncertainty (2019, methodology): “For direct comparison with [19], we follow exactly the same setting: Market-1501, CUHK01, and CUHK03 are used with the same splits as in [19] (see Supplementary Material for details).”
• Multi-scale Deep Learning Architectures for Person Re-identification (2017, methodology): “CUHK03 [27], CUHK01 [28] and VIPeR [12].”
• Learning Mid-level Filters for Person Re-identification (2014, methodology): “Existing person re-identification approaches work on three aspects: distance learning [29, 4, 12, 18, 30, 16, 14], feature design and selection [5, 3, 17, 15, 28, 7, 20, 25], and mid-level feature learning [11, 22, 10, 13].”
• Patch-based Knowledge Distillation for Lifelong Person Re-Identification (2022, background): “It consists of twelve ReID datasets, including five seen datasets(Market-1501[73],CUHK-SYSU[64],DukeMTMC-reID[48], MSMT17_V2[61]andCUHK03[32])fornon-forgettingevaluation andsevenunseendatasets(VIPeR[16],PRID[22],GRID[39],i-LIDS[74],CUHK01[31],CUHK02[30]andSenseReID[70])for generalizationevaluation.”
• Person Re-identification: Past, Present and Future (2016, methodology): “A number of datasets for image-based re-ID have been released, and some commonly used datasets are summarized in Table 3 [16], [36], [44], [82], [83], [84], [85], [86], [87], [88], [89], [90] for image-based re-ID. and testing sets.”
• Human Semantic Parsing for Person Re-identification (2018, methodology): “In addition to these datasets that were used in evaluation, we utilize 3DPeS [4], CUHK01 [23], CUHK02 [22], PRID [17], PSDB [41], Shinpuhkan [20] and VIPeR [15] datasets to augment our training data.”