Multi-Granularity Class Prototype Topology Distillation for Class-Incremental Source-Free Unsupervised Domain Adaptation

Abstract

This paper explores the Class-Incremental Source-Free Unsupervised Domain Adaptation~(CI-SFUDA) problem, where the unlabeled target data come incrementally without access to labeled source instances. This problem poses two challenges, the interference of similar source-class knowledge in target-class representation learning and the shocks of new target knowledge to old ones. To address them, we propose the Multi-Granularity Class Prototype Topology Distillation~(GROTO) algorithm, which effectively transfers the source knowledge to the class-incremental target domain. Concretely, we design the multi-granularity class prototype self-organization module and the prototype topology distillation module. First, we mine the positive classes by modeling accumulation distributions. Next, we introduce multi-granularity class prototypes to generate reliable pseudo-labels, and exploit them to promote the positive-class target feature self-organization. Second, the positive-class prototypes are leveraged to construct the topological structures of source and target feature spaces. Then, we perform the topology distillation to continually mitigate the shocks of new target knowledge to old ones. Extensive experiments demonstrate that our proposed method achieves state-of-the-art performance on three public datasets.

🔧 Installation

• Clone this repository:

git clone https://github.com/dengpeihua/GROTO.git
cd GROTO

• Install the environment by running the following command:

conda env create -f environment.yml

📊 Data Preparation

The files of data list and their corresponding labels have been put in the directory ./data_splits, and the imagenet_list.txt can be downloaded at GoogleDrive.

Please manually download the office31, office-home and ImageNet-Caltech benchmarks from here, here and here, then put them into the corresponding directory (e.g., './dataset/office-home').

Put the corresponding file in your path (e.g., './dataset/office-home/Art.txt')

⚡ Source Pre-trained

To obtain the pre-trained model on the source domain:

• Pre-train on the amazon domain of Office-31-CI:

python source_train_office31.py --gpu 0 --source 0

• Pre-train on the Art domain of Office-Home-CI:

python source_train_office-home.py --gpu 0 --source 0

• from Caltech256 to ImageNet84:

python source_train_cal256.py --gpu 0

• from ImageNet1K to Caltech84:

python source_train_imagenet1k.py --gpu 0

➡️ Class-Incremental Source-Free Unsupervised Domain Adaptation

Run .sh scripts in the sh folder to obtain the results of different datasets.

💡 Checkpoints & Logs

We provide all source domain pre-trained models of three datasets, as well as the checkpoints and logs after class-incremental source-free unsupervised domain adaptation at GoogleDrive.

🙏 Acknowledgments

We would like to thank the authors of previous related projects for generously sharing their code and insights: ViT, and ProCA.

🤝 Citation

If you find our work useful, please consider giving a citation 📝 :)

@InProceedings{Deng_2025_CVPR,
    author    = {Deng, Peihua and Zhang, Jiehua and Sheng, Xichun and Yan, Chenggang and Sun, Yaoqi and Fu, Ying and Li, Liang},
    title     = {Multi-Granularity Class Prototype Topology Distillation for Class-Incremental Source-Free Unsupervised Domain Adaptation},
    booktitle = {Proceedings of the Computer Vision and Pattern Recognition Conference (CVPR)},
    month     = {June},
    year      = {2025},
    pages     = {30566-30576}
}