Learning Source-free Domain Adaptation for Visible-Infrared Person Re-Identification (NeruIPS, PyTorch Code)
Authors: Yongxiang Li, Yanglin Feng, Yuan Sun, Dezhong Peng, Xi Peng and Peng Hu
In this paper, we investigate source-free domain adaptation (SFDA) for visible-infrared person re-identification (VI-ReID), aiming to adapt a pre-trained source model to an unlabeled target domain without access to source data. To address this challenging setting, we propose a novel learning paradigm, termed Source-Free Visible-Infrared Person Re-Identification (SVIP), which fully exploits the prior knowledge embedded in the source model to guide target domain adaptation. The proposed framework comprises three key components specifically designed for the source-free scenario: 1) a Source-Guided Contrastive Learning (SGCL) module, which leverages the discriminative feature space of the frozen source model as a reference to perform contrastive learning on the unlabeled target data, thereby preserving discrimination without requiring source samples; 2) a Residual Transfer Learning (RTL) module, which learns residual mappings to adapt the target model’s representations while maintaining the knowledge from the source model; and 3) a Structural Consistency-Guided Cross-modal Alignment (SCCA) module, which enforces reciprocal structural constraints between visible and infrared modalities to identify reliable cross-modal pairs and achieve robust modality alignment without source supervision. Extensive experiments on benchmark datasets demonstrate that SVIP substantially enhances target domain performance and outperforms existing unsupervised VI-ReID methods under source-free settings.
In the following instructions, we take the transfer setting SYSU-MM01 → RegDB as a typical example.
Please change the dataset path to your own path in the prepare_regdb.py.
python prepare_regdb.pyPut your source model trained on SYSU-MM01 into ./pretrained and name it resnet50-sysumm01.pth.
bash run_train_regdb.shbash test_regdb.shTo evaluate under the weather degradation setting, generate the required traget domain data with the following command. The example images in ./data_corruptions/examples will be transformed and saved to ./data_corruptions/examples_results.
python make_corrupt.pyIf our work is helpful for your research, please consider citing:
@inproceedings{li2025learning,
title={Learning Source-Free Domain Adaptation for Visible-Infrared Person Re-Identification},
author={Li, Yongxiang and Feng, Yanglin and Sun, Yuan and Peng, Dezhong and Peng, Xi and Hu, Peng},
booktitle={The Thirty-ninth Annual Conference on Neural Information Processing Systems (NeurIPS)},
year={2025}
}