Exploit the Unknown Gradually: One-Shot Video-Based Person Re-identification by Stepwise Learning

By Yuehua Wu, Yutian Lin, Xuanyi Dong, Yan Yan, Wanli Ouyang, Yi Yang. 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2018

377 citations recorded by the source corpus.

We focus on the one-shot learning for video-based person re-Identification (re-ID). Unlabeled tracklets for the person re-ID tasks can be easily obtained by preprocessing, such as pedestrian detection and tracking. In this paper, we propose an approach to exploiting unlabeled tracklets by gradually but steadily improving the discriminative capability of the Convolutional Neural Network (CNN) feature representation via stepwise learning. We first initialize a CNN model using one labeled tracklet for each identity. Then we update the CNN model by the following two steps iteratively: 1. sample a few candidates with most reliable pseudo labels from unlabeled tracklets; 2. update the CNN model according to the selected data. Instead of the static sampling strategy applied in existing works, we propose a progressive sampling method to increase the number of the selected pseudo-labeled candidates step by step. We systematically investigate the way how we should select pseudo-labeled tracklets into the training set to make the best use of them. Notably, the rank-1 accuracy of our method outperforms the state-of-the-art method by 21.46 points (absolute, i.e., 62.67% vs. 41.21%) on the MARS dataset, and 16.53 points on the DukeMTMC-VideoReID dataset1.

Subfields: Computer Vision and Pattern Recognition, Safety, Risk, Reliability and Quality

View Paper: https://www.semanticscholar.org/paper/b83964876faa536489b4c189fdeb1f9d50172862

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. Liang Zheng
3. Luc Van Gool
4. Dacheng Tao
5. Wanli Ouyang
6. Xiaogang Wang
7. Yunchao Wei
8. Ling Shao

In Safety, Risk, Reliability and Quality

1. Liang Zheng
2. Yi Yang
3. Xiaojun Chang
4. Xuanyi Dong

Citations in the corpus

• Improving Person Re-identification by Attribute and Identity Learning (2017, methodology): “Recently, some semi-supervised methods [37, 19] and unsupervised methods [38, 22] has been proposed to address the data problem for re-ID.”
• Robust Anchor Embedding for Unsupervised Video Person re-IDentification in the Wild (2018, background): “Moreover, most of these methods suffer from the scalability issues, thus cannot be applied to large-scale applications [9, 43, 53].”
• Image-Image Domain Adaptation with Preserved Self-Similarity and Domain-Dissimilarity for Person Re-identification (2017, background): “[46] propose a dynamic sampling stragy for one-shot video-based re-ID.”
• Unsupervised Person Re-identification (2017, methodology): “In this paper, we utilize SPL to select reliable samples to fine tune the original model.”
• 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].”
• Pose-Guided Feature Alignment for Occluded Person Re-Identification (2019, background): “Recent years, deep learning methods [39, 34, 13, 27, 3, 25, 7, 6, 36, 21] for person re-id show significant superiority on retrieval accuracy.”
• A Survey of Deep Active Learning (2020, methodology): “[4, 230] PRID [93], MARS [256], BDD100K [246], DukeMTMC-VideoReID [237], CityPersons [249], Caltech Pedestrian[52] Person Re-id”
• A Comprehensive Survey of Few-shot Learning: Evolution, Applications, Challenges, and Opportunities (2022, methodology): “Speciically, Wu et al [63] selected labeled data as the initial model, Wang et al.”
• Style Aggregated Network for Facial Landmark Detection (2018, background): “These mean face images show the very similar face, while they have quite different environments. taking two complementary face images as the input benefits the facial landmark prediction results.”
• Continual egocentric object recognition (2019, methodology): “The most prominent applications in this field are related to the iden-tification of humans [35], with some works focusing on vehicle re-identification [29].”
• Self-Similarity Grouping: A Simple Unsupervised Cross Domain Adaptation Approach for Person Re-Identification (2018, methodology): “[41] propose a progressive sampling method to gradually predict reliable pseudo labels and update deep model for one shot video-based re-ID.”
• One Venue, Two Conferences: The Separation of Chinese and American Citation Networks (2022, methodology): “Furthermore, researchers in China have continued to use the Duke dataset even after its removal [17, 7, 19].”
• A Bottom-Up Clustering Approach to Unsupervised Person Re-Identification (2019, methodology): “As discussed in (Wu et al. 2018a), these approaches are actually the one-example methods.”
• Generalizing a Person Retrieval Model Hetero- and Homogeneously (2018, methodology): “Benefit from the remarkable success of deep learning [24,25,26,27,28], recent works [2,29,30] attempt to predict pseudo-labels to unlabeled samples based on the deep learning framework.”
• Generalizing from a Few Examples (2019, methodology): “In [148], a progressive strategy is used to select informative unlabeled samples.”