Source-Free Domain Adaptation (SFDA) aims to train a target model without source data, and the key is to generate pseudo-labels using a pre-trained source model. However, we observe that the source model often produces highly uncertain pseudo-labels for hard samples, particularly those heavily affected by domain shifts, leading to these noisy pseudo-labels being introduced even before adaptation and further reinforced through parameter updates. Additionally, they continuously influence neighbor samples through propagation in the featurethis http URLeliminate the issue of noise accumulation, we propose a novel Progressive Curriculum Labeling (ElimPCL) method, which iteratively filters trustworthy pseudo-labeled samples based on prototype consistency to exclude high-noise samples from training. Furthermore, a Dual MixUP technique is designed in the feature space to enhance the separability of hard samples, thereby mitigating the interference of noisy samples on theirthis http URLexperiments validate the effectiveness of ElimPCL, achieving up to a 3.4% improvement on challenging tasks compared to state-of-the-art methods.
View on arXiv@article{cheng2025_2503.23712,
title={ ElimPCL: Eliminating Noise Accumulation with Progressive Curriculum Labeling for Source-Free Domain Adaptation },
author={ Jie Cheng and Hao Zheng and Meiguang Zheng and Lei Wang and Hao Wu and Jian Zhang },
journal={arXiv preprint arXiv:2503.23712},
year={ 2025 }
}