Dynamic Feature Pruning and Consolidation for Occluded Person Re-identification
Authors: YuTeng Ye, Hang Zhou, Jiale Cai, Chenxing Gao, Youjia Zhang, Junle Wang, Qiang Hu, Junqing Yu, Wei Yang
AAAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experimental results demonstrate the effectiveness of our proposed framework on occluded, partial, and holistic Re-ID datasets. In particular, our method outperforms state-of-the-art results by at least 8.6% m AP and 6.0% Rank-1 accuracy on the challenging Occluded-Duke dataset. |
| Researcher Affiliation | Collaboration | Yu Teng Ye1, Hang Zhou1, Jiale Cai1, Chenxing Gao1, Youjia Zhang1, Junle Wang2, Qiang Hu3, Junqing Yu1, Wei Yang1* 1Huazhong University of Science and Technology, Wuhan, China 2Tencent 3Shanghai Jiao Tong University, Shanghai, China |
| Pseudocode | No | The paper describes the proposed framework using natural language and diagrams, but no formal pseudocode or algorithm blocks are provided. |
| Open Source Code | No | The paper does not provide a specific link or explicit statement about the availability of the source code for the described methodology. |
| Open Datasets | Yes | Occluded-Duke (Miao et al. 2019) includes 15,618 training images of 702 persons, 2,210 occluded query images of 519 persons, and 17,661 gallery images of 1,110 persons. Occluded-Re ID (Zhuo et al. 2018) consists of 1,000 occluded query images and 1,000 full-body gallery images both belonging to 200 identities. Partial-Re ID (Zheng et al. 2015b) involves 600 images from 60 persons, and each person consists 5 partial and 5 full-body images. Market1501 (Zheng et al. 2015a) contains 12,936 training images of 751 persons, 19,732 query images and 3,368 gallery images of 750 persons captured by six cameras. |
| Dataset Splits | Yes | Occluded-Duke (Miao et al. 2019) includes 15,618 training images of 702 persons, 2,210 occluded query images of 519 persons, and 17,661 gallery images of 1,110 persons. Market1501 (Zheng et al. 2015a) contains 12,936 training images of 751 persons, 19,732 query images and 3,368 gallery images of 750 persons captured by six cameras. |
| Hardware Specification | No | The Acknowledgments section mentions: "The computation is completed in the HPC Platform of Huazhong University of Science and Technology." However, no specific details about GPU models, CPU models, or other hardware specifications are provided. |
| Software Dependencies | No | The paper mentions using "Vi T-B/16 as our backbone" but does not specify other software dependencies like PyTorch, TensorFlow, or CUDA versions with specific numbers. |
| Experiment Setup | Yes | In the training process, we resize all input images to 256 128. The training images are augmented with random horizontal flipping, padding, random cropping and random erasing (Zhong et al. 2020). The batch size is 64 with 4 images per ID and the learning rate is initialized as 0.008 with cosine learning rate decay. The distance weight α in eq. (10) is 0.4 and the parameter of keep rate γ is 0.8. |