NTIRE 2025 Challenge on Real-World Face Restoration: Methods and Results
Zheng Chen, Jingkai Wang, Kai Liu, Jue Gong, Lei Sun, Zongwei Wu, Radu Timofte, Yulun Zhang, Jianxing Zhang, Jinlong Wu, Jun Wang, Zheng Xie, Hakjae Jeon, Suejin Han, Hyung-Ju Chun, Hyunhee Park, Zhicun Yin, Junjie Chen, Ming Liu, Xiaoming Li, Chao Zhou, Wangmeng Zuo, Weixia Zhang, Dingquan Li, Kede Ma, Yun Zhang, Zhuofan Zheng, Yuyue Liu, Shizhen Tang, Zihao Zhang, Yi Ning, Hao Jiang, Wenjie An, Kangmeng Yu, Chenyang Wang, Kui Jiang, Xianming Liu, Junjun Jiang, Yingfu Zhang, Gang He, Siqi Wang, Kepeng Xu, Zhenyang Liu, Changxin Zhou, Shanlan Shen, Yubo Duan, Yiang Chen, Jin Guo, Mengru Yang, Jen-Wei Lee, Chia-Ming Lee, Chih-Chung Hsu, Hu Peng, Chunming He
TL;DR
The NTIRE 2025 Real-world Face Restoration challenge tackles recovering HQ, identity-consistent faces from diverse real-world degradations without resource limits. A broad set of approaches emerged, led by diffusion-based priors and Transformer priors, often in multi-stage pipelines that fuse priors and perform per-image refinement. Evaluation combines AdaFace identity checks with a suite of no-reference IQA metrics (CLIPIQA, MANIQA, MUSIQ, Q-Align), contextualized by FFHQ-based training and strict test-disjointness to ensure fair comparisons. Top teams demonstrated that integrating generative priors (e.g., StyleGAN, SDXL-Turbo), discriminative cues, and image-specific fine-tuning yields natural, identity-preserving restorations and pushes real-world face restoration toward deployment-ready performance.
Abstract
This paper provides a review of the NTIRE 2025 challenge on real-world face restoration, highlighting the proposed solutions and the resulting outcomes. The challenge focuses on generating natural, realistic outputs while maintaining identity consistency. Its goal is to advance state-of-the-art solutions for perceptual quality and realism, without imposing constraints on computational resources or training data. The track of the challenge evaluates performance using a weighted image quality assessment (IQA) score and employs the AdaFace model as an identity checker. The competition attracted 141 registrants, with 13 teams submitting valid models, and ultimately, 10 teams achieved a valid score in the final ranking. This collaborative effort advances the performance of real-world face restoration while offering an in-depth overview of the latest trends in the field.