Cross-Modality Paired-Images Generation for RGB-Infrared Person Re-Identification
Authors: Guan-An Wang, Tianzhu Zhang, Yang Yang, Jian Cheng, Jianlong Chang, Xu Liang, Zeng-Guang Hou12144-12151
AAAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Extensive experimental results on two standard benchmarks demonstrate that the proposed model favourably against state-of-the-art methods. |
| Researcher Affiliation | Academia | 1Institute of Automation, Chinese Academy of Sciences, Beijing, China 2University of Chinese Academy of Sciences, Beijing, China 3Center for Excellence in Brain Science and Intelligence Technology, Beijing, China 4University of Science and Technology of China, Beijing, China {wangguanan2015, liangxu2013, zengguang.hou}@ia.ac.cn, tzzhang@ustc.edu.cn, {yang.yang, jcheng, jianlong.chang}@nlpr.ia.ac.cn |
| Pseudocode | No | The paper describes its methodology in text and through architectural diagrams (e.g., Figure 3), but it does not include a formally labeled 'Pseudocode' or 'Algorithm' block. |
| Open Source Code | Yes | Code is available at https://github.com/wangguanan/JSIA-Re ID. |
| Open Datasets | Yes | We evaluate our model on two standard benchmarks including SYSU-MM01 and Reg DB. (1) SYSU-MM01 (Wu et al. 2017) is a popular RGB-IR Re-ID dataset... (2) Reg DB (Nguyen et al. 2017) contains 412 persons... |
| Dataset Splits | No | For SYSU-MM01: 'The training set contains 19,659 RGB images and 12,792 IR images of 395 persons and the test set contains 96 persons.' For Reg DB: 'the results of Reg DB are based on the average of 10 times repeated random split of training and testing sets.' The paper specifies training and testing sets but does not explicitly detail a separate validation set. |
| Hardware Specification | No | The paper does not specify any particular hardware components such as GPU or CPU models used for running the experiments. |
| Software Dependencies | No | The paper mentions using 'Res Net-50' and 'LSGAN' as components, but it does not specify software dependencies with version numbers (e.g., Python, PyTorch, TensorFlow, or specific library versions). |
| Experiment Setup | Yes | Thus, the overall objective function of our method can formulated as below: L = λcyc Lcyc + λgan Lgan + λalign Lalign + λreid(Lcls + Ltriplet) (11) where λ are weights of corresponding terms. Following (Zhu et al. 2017), we set λcyc = 10 and λgan = 1. λreid is set 1 empirically and λalign is decided by grid search. ... In generation module G, following (Radford, Metz, and Chintala 2016), we construct our modality-specific encoders with 2 strided convolutional layers followed by a global average pooling layer and a fully connected layer. For decoders, following (Wang et al. 2017), we use 4 residual blocks with Adaptive Instance Normalization (Ada IN) and 2 upsampling with convolutional layers. |