Diffusion Prior Interpolation for Flexibility Real-World Face Super-Resolution
Jiarui Yang, Tao Dai, Yufei Zhu, Naiqi Li, Jinmin Li, Shu-Tao Xia
Abstract
Diffusion models represent the state-of-the-art in generative modeling. Due to their high training costs, many works leverage pre-trained diffusion models' powerful representations for downstream tasks, such as face super-resolution (FSR), through fine-tuning or prior-based methods. However, relying solely on priors without supervised training makes it challenging to meet the pixel-level accuracy requirements of discrimination task. Although prior-based methods can achieve high fidelity and high-quality results, ensuring consistency remains a significant challenge. In this paper, we propose a masking strategy with strong and weak constraints and iterative refinement for real-world FSR, termed Diffusion Prior Interpolation (DPI). We introduce conditions and constraints on consistency by masking different sampling stages based on the structural characteristics of the face. Furthermore, we propose a condition Corrector (CRT) to establish a reciprocal posterior sampling process. DPI can balance consistency and diversity and can be seamlessly integrated into pre-trained models. In extensive experiments conducted on synthetic and real datasets, along with consistency validation in face recognition, DPI demonstrates superiority over SOTA FSR methods.
BibTeX
@article{Yang_Dai_Zhu_Li_Li_Xia_2025, title={Diffusion Prior Interpolation for Flexibility Real-World Face Super-Resolution}, volume={39}, url={https://ojs.aaai.org/index.php/AAAI/article/view/32997}, DOI={10.1609/aaai.v39i9.32997}, abstractNote={Diffusion models represent the state-of-the-art in generative modeling. Due to their high training costs, many works leverage pre-trained diffusion models’ powerful representations for downstream tasks, such as face super-resolution (FSR), through fine-tuning or prior-based methods. However, relying solely on priors without supervised training makes it challenging to meet the pixel-level accuracy requirements of discrimination task. Although prior-based methods can achieve high fidelity and high-quality results, ensuring consistency remains a significant challenge. In this paper, we propose a masking strategy with strong and weak constraints and iterative refinement for real-world FSR, termed Diffusion Prior Interpolation (DPI). We introduce conditions and constraints on consistency by masking different sampling stages based on the structural characteristics of the face. Furthermore, we propose a condition Corrector (CRT) to establish a reciprocal posterior sampling process. DPI can balance consistency and diversity and can be seamlessly integrated into pre-trained models. In extensive experiments conducted on synthetic and real datasets, along with consistency validation in face recognition, DPI demonstrates superiority over SOTA FSR methods.}, number={9}, journal={Proceedings of the AAAI Conference on Artificial Intelligence}, author={Yang, Jiarui and Dai, Tao and Zhu, Yufei and Li, Naiqi and Li, Jinmin and Xia, Shu-Tao}, year={2025}, month={Apr.}, pages={9211-9219} }