This paper introduces a novel approach to leverage the generalizability capability of Diffusion Models for Source-Free Domain Adaptation (DM-SFDA). Our proposed DM-SFDA method involves fine-tuning a pre-trained text-to-image diffusion model to generate source domain images using features from the target images to guide the diffusion process. Specifically, the pre-trained diffusion model is fine-tuned to generate source samples that minimize entropy and maximize confidence for the pre-trained source model. We then apply established unsupervised domain adaptation techniques to align the generated source images with target domain data. We validate our approach through comprehensive experiments across a range of datasets, including Office-31, Office-Home, and VisDA. The results highlight significant improvements in SFDA performance, showcasing the potential of diffusion models in generating contextually relevant, domain-specific images.

本文介绍了一种新的方法，利用扩散模型(Diffusion Models)对源无关域适应(Source-Free Domain Adaptation, SFDA)的一般化能力。我们提出的DM-SFDA方法包括对预训练的文本-图像扩散模型进行微调，以利用目标图像特征来指导扩散过程，生成源域图像。具体来说，预训练的扩散模型通过微调生成能够在预训练的源模型中最小化熵并最大化置信度的源样本。然后，我们应用已建立的无监督域适应技术来将生成的源图像与目标域数据对齐。我们通过包括Office-31、Office-Home和VisDA在内的多个数据集进行了全面实验验证。实验结果突出了SFDA性能的显著提升，展示了扩散模型在生成与上下文相关的域特定图像方面的潜力。

无源领域适应与扩散指导的源数据生成