Pedestrian Attribute Recognition as Label-balanced Multi-label Learning
Authors: Yibo Zhou, Hai-Miao Hu, Yirong Xiang, Xiaokang Zhang, Haotian Wu
ICML 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We conduct a thorough evaluation of our method, comparing with strong baselines and a range of recent approaches. The results are presented in Table 1. |
| Researcher Affiliation | Academia | 1Beijing Key Laboratory of Digital Media, School of Computer Science and Engineering, Beihang University, Beijing 100191, China 2State Key Laboratory of Virtual Reality Technology and Systems, Beihang University, Beijing 100191, China 3Hangzhou Innovation Institute of Beihang University, Hangzhou 310051, China 4The University of Manchester, UK. |
| Pseudocode | Yes | Algorithm 1 Pseudo-code of the GOAT-enhanced FRDL |
| Open Source Code | Yes | Other details can be referred in our code at github. |
| Open Datasets | Yes | We perform experiments on popular large-scale PAR datasets of PA100k (Liu et al., 2017), PETA (Deng et al., 2014) and RAPv1 (Li et al., 2016). |
| Dataset Splits | Yes | For the datasets configuration, we strictly follow (Jia et al., 2020) to make a wide and fair comparison with prior arts. |
| Hardware Specification | No | The paper does not provide specific hardware details such as GPU or CPU models used for running experiments. |
| Software Dependencies | No | The paper mentions 'Adam solver is applied' but does not specify software names with version numbers (e.g., Python, PyTorch, TensorFlow versions, or other libraries). |
| Experiment Setup | Yes | Image is spatially resized to 256 192 for input, and batch size is set as 64. Adam solver is applied with weight decay of 5e-4. Horizontal flip and random crop are the only image augmentation methods. The learning rate starts at 1e-4 and decays by a factor of 10 at certain steps. Unless otherwise stated, we on default set the T in Algorithm 1 as 20. |