Self-Guided Hard Negative Generation for Unsupervised Person Re-Identification
Authors: Dongdong Li, Zhigang Wang, Jian Wang, Xinyu Zhang, Errui Ding, Jingdong Wang, Zhaoxiang Zhang
IJCAI 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We conduct experiments on three widely-used person re-ID datasets, i.e. Market-1501 [Zheng et al., 2015], Duke MTMCRe ID [Zheng et al., 2017] and MSMT17 [Wei et al., 2018] to validate our method. |
| Researcher Affiliation | Collaboration | Dongdong Li1,3 , Zhigang Wang2 , Jian Wang2 , Xinyu Zhang2 , Errui Ding2 , Jingdong Wang2 , Zhaoxiang Zhang1,3,4 1Institute of Automation, Chinese Academy of Sciences (CASIA), China 2Baidu VIS, China 3University of Chinese Academy of Sciences (UCAS), China 4Centre for Artificial Intelligence and Robotics, HKISI CAS, China |
| Pseudocode | No | The paper does not contain structured pseudocode or algorithm blocks. |
| Open Source Code | No | The paper states 'We implement our method mainly based on the open-source codes released by [Ge et al., 2020] and [Chen et al., 2021a].' This indicates the use of *other* open-source code, not that the authors' own code for this paper is openly available. No specific repository link or explicit code release statement for their work is provided. |
| Open Datasets | Yes | We conduct experiments on three widely-used person re-ID datasets, i.e. Market-1501 [Zheng et al., 2015], Duke MTMCRe ID [Zheng et al., 2017] and MSMT17 [Wei et al., 2018] to validate our method. |
| Dataset Splits | Yes | Market-1501 consists of 32,668 images of 1501 persons captured by 6 cameras, where 12,936 images of 751 persons constitute the training set. The remaining images are for test. Duke MTMC-Re ID contains 36,411 images of 1,404 persons from 8 cameras in total. ... The training set is comprised of 16,522 images of 702 persons. Other images constitute the test set. MSMT17 is a challenging dataset containing 126,441 images of 4,101 persons. These images are captured by 15 cameras including 32,621 training images of 1,041 persons and 93,820 test images of 3,060 persons. |
| Hardware Specification | Yes | The extra training procedure can be completed in a few hours by modern GPUs, e.g. NVIDIA P40. |
| Software Dependencies | No | The paper mentions software components like 'ResNet50' (model architecture), 'Adam optimizer' (algorithm), and 'DBSCAN' (clustering algorithm) but does not provide specific version numbers for any software libraries, frameworks, or programming languages used. |
| Experiment Setup | Yes | During training, images are resized to 256 128. Random flip, random erasing and image padding are adopted as data augmentations. ... The learning rate is initialized as 1 10 4 for Adam, and the weight decay is set as 5 10 4. ... The temperature τ in Eq. 4 and Eq. 7 is set to 0.05. The widely-used DBSCAN is employed as the clustering algorithm in our method. ... we use identity-based sampling strategy in our training, which first randomly selects 16 identities and further randomly selects 4 instances for each identity to constitute a mini-batch of size 64. ... For the former, we train the re-ID model following the settings of a basic unsupervised re-ID method. Besides, we train the generator with only reconstruction loss for 50 epochs, where the learning rate is multiplied by 0.1 after 20 epochs. For the joint fine-tune stage, the whole framework is fine-tuned for 20 epochs, where the learning rate is multiplied by 0.1 after 10 epochs. |