End-to-End Deep Learning for Person Search

By Tong Xiao, Shuang Li, Bochao Wang, Liang Lin, Xiaogang Wang. 2016

238 citations recorded by the source corpus.

How researchers see this paper

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

1. Liang Zheng
2. Yi Yang
3. Zhedong Zheng
4. Xiaogang Wang
5. Wanli Ouyang
6. Tao A. Xiang
7. Mang Ye
8. Dacheng Tao

Citations in the corpus

• A Discriminatively Learned CNN Embedding for Person Reidentification (2016, methodology): “In an attempt to take full advantages of these labels, classification models which treat person re-identification as a multi-class recognition task, are employed for feature lea rning [29, 35, 30].”
• Person Re-identification: Past, Present and Future (2016, methodology): “This work views it as an “ultimate” goal to integrate detection, tracking, and retrieval into one framework, and evaluate the impact of each module on the [18], [77], [128], [133] for end-to-end person re-identification (search). overall re-ID performance.”
• See the Forest for the Trees: Joint Spatial and Temporal Recurrent Neural Networks for Video-Based Person Re-identification (2017, unknown): “A large portion of existing works lies in the former category [7, 21, 28, 32, 33, 34] while only a few belong to the latter.”
• Learning Continual Compatible Representation for Re-indexing Free Lifelong Person Re-identification (2024, methodology): “To verify the effectiveness of our method, we conduct extensive experiments on five benchmark lifelong person ReID datasets, including Market-1501 [41], CUHK-SYSU [36], DukeMTMC-ReID [43], MSMT17-V2 [34] and CUHK03 [12].”
• Lifelong Person Re-Identification via Adaptive Knowledge Accumulation (2021, methodology): “Specifically, we selected five relatively large-scale datasets, CUHK03 (CU) [18], Market-1501 (MA) [52], MSMT17 V2 (MS) [40], DukeMTMC-ReID (DU) [54] and CUHK-SYSU ReID (SY) [43], and sampled 500 identities from each of their training sets to construct five training domains so that each domain has a”
• Person Search with Natural Language Description (2017, methodology): “We collected 40,206 images of 13,003 persons from five existing person re-identification datasets, CUHK03 [22], Market-1501 [42], SSM [38], VIPER [9], and CUHK01 [21], as the subjects for language descriptions.”
• Improving Person Re-identification by Attribute and Identity Learning (2017, background): “ies may be compromised. The second category, i.e., representation learning, has gained increasing popularity because it yields superior accuracy [52] and does not harm the efficiency. Examples include [41,49,42,9,53]. Xiao et al. [41] propose to learn a generic feature embedding by training a classifi”
• 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.”
• Patch-based Knowledge Distillation for Lifelong Person Re-Identification (2022, background): “NotethatCUHK-SYSUisapersonsearch dataset,sowefollowtheprocedurein[46]byfirstcroppingthe imagesusingtheground-truthpersonboundingboxannotation andthenselectingasubsetofidentitieswithmorethan4bounding boxes.”
• Multi-scale Deep Learning Architectures for Person Re-identification (2017, methodology): “In contrast, this paper formulates person re-id into two tasks: classification [55, 56] and verification [1, 48].”
• 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.”
• Part-Aligned Bilinear Representations for Person Re-identification (2018, methodology): “This problem causes conventional global representations [2–7] or strip/grid-based representations [8–13] to be unreliable as they implicitly assume that every person appears in a similar pose within a tightly surrounded bounding box.”
• Learning Deep Context-Aware Features over Body and Latent Parts for Person Re-identification (2017, background): “With the increasing sample size of ReID dataset, the learning of features from multi-class person identification tasks [30, 39, 40, 52, 55], denoted as ID-discriminative Embedding (IDE) [55], has shown great potentials on current large-scale person ReID datasets, such as MARS [52] and PRW [55], wher”
• Style Normalization and Restitution for Generalizable Person Re-Identification (2020, methodology): “We follow [36] and train SNR on the same five datasets (M+D+C+CUHK02[23]+CUHK-SYSU[47]).”
• Spindle Net: Person Re-identification with Human Body Region Guided Feature Decomposition and Fusion (2017, methodology): “The CUHK02 [12] and PSDB [30] datasets are also used for training but not evaluated.”
• Unlabeled Samples Generated by GAN Improve the Person Re-identification Baseline in Vitro (2017, unknown): “The code is available at https://github.com/layumi/Person-reID_GAN .”