The Second Challenge on Real-World Face Restoration at NTIRE 2026: Methods and Results

Jingkai Wang, Jue Gong, Zheng Chen, Kai Liu, Jiatong Li, Yulun Zhang, Radu Timofte, Jiachen Tu, Yaokun Shi, Guoyi Xu, Yaoxin Jiang, Jiajia Liu, Yingsi Chen, Yijiao Liu, Hui Li, Yu Wang, Congchao Zhu, Alexandru-Gabriel Lefterache, Anamaria Radoi, Chuanyue Yan, Tao Lu, Yanduo Zhang, Kanghui Zhao, Jiaming Wang, Yuqi Li, WenBo Xiong, Yifei Chen, Xian Hu, Wei Deng, Daiguo Zhou, Sujith Roy V, Claudia Jesuraj, Vikas B, Spoorthi LC, Nikhil Akalwadi, Ramesh Ashok Tabib, Uma Mudenagudi, Yuxuan Jiang, Chengxi Zeng, Tianhao Peng, Fan Zhang, David Bull Wei Zhou, Linfeng Li, Hongyu Huang, Hoyoung Lee, SangYun Oh, ChangYoung Jeong, Axi Niu, Jinyang Zhang, Zhenguo Wu, Senyan Qing, Jinqiu Sun, Yanning Zhang · 2026

Read The Second Challenge on Real-World Face Restoration at NTIRE 2026: Methods and Results by Jingkai Wang, Jue Gong, Zheng Chen, Kai Liu, Jiatong Li, Yulun Zhang, Radu Timofte, Jiachen Tu, Yaokun Shi, Guoyi Xu, Yaoxin Jiang, Jiajia Liu, Yingsi Chen, Yijiao Liu, Hui Li, Yu Wang, Congchao Zhu, Alexandru-Gabriel Lefterache, Anamaria Radoi, Chuanyue Yan, Tao Lu, Yanduo Zhang, Kanghui Zhao, Jiaming Wang, Yuqi Li, WenBo Xiong, Yifei Chen, Xian Hu, Wei Deng, Daiguo Zhou, Sujith Roy V, Claudia Jesuraj, Vikas B, Spoorthi LC, Nikhil Akalwadi, Ramesh Ashok Tabib, Uma Mudenagudi, Yuxuan Jiang, Chengxi Zeng, Tianhao Peng, Fan Zhang, David Bull Wei Zhou, Linfeng Li, Hongyu Huang, Hoyoung Lee, SangYun Oh, ChangYoung Jeong, Axi Niu, Jinyang Zhang, Zhenguo Wu, Senyan Qing, Jinqiu Sun, Yanning Zhang (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-10532.txt.

Introduction

Wang and 52 colleagues review the second NTIRE face restoration contest. The task prizes natural looks that keep the same face, with no caps on compute or training data. Scores blend image quality with an AdaFace identity check. Of 96 entrants, 10 teams filed valid models and 9 reached the final board. The review maps the top tricks and trends shaping real world face repair.

Catalog record

CREATORJingkai Wang, Jue Gong, Zheng Chen, Kai Liu, Jiatong Li, Yulun Zhang, Radu Timofte, Jiachen Tu, Yaokun Shi, Guoyi Xu, Yaoxin Jiang, Jiajia Liu, Yingsi Chen, Yijiao Liu, Hui Li, Yu Wang, Congchao Zhu, Alexandru-Gabriel Lefterache, Anamaria Radoi, Chuanyue Yan, Tao Lu, Yanduo Zhang, Kanghui Zhao, Jiaming Wang, Yuqi Li, WenBo Xiong, Yifei Chen, Xian Hu, Wei Deng, Daiguo Zhou, Sujith Roy V, Claudia Jesuraj, Vikas B, Spoorthi LC, Nikhil Akalwadi, Ramesh Ashok Tabib, Uma Mudenagudi, Yuxuan Jiang, Chengxi Zeng, Tianhao Peng, Fan Zhang, David Bull Wei Zhou, Linfeng Li, Hongyu Huang, Hoyoung Lee, SangYun Oh, ChangYoung Jeong, Axi Niu, Jinyang Zhang, Zhenguo Wu, Senyan Qing, Jinqiu Sun, Yanning Zhang
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE36958
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.10532

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

The Second Challenge on Real-World Face Restoration at NTIRE 2026: Methods and Results Jingkai Wang, Jue Gong, Zheng Chen, Kai Liu, Jiatong Li, Yulun Zhang, Radu Timofte, Jiachen Tu, Yaokun Shi, Guoyi Xu, Yaoxin Jiang, Jiajia Liu, Yingsi Chen, Yijiao Liu, Hui Li, Yu Wang, Congchao Zhu, Alexandru-Gabriel Lefterache, Anamaria Radoi, Chuanyue Yan, Tao Lu, Yanduo Zhang, Kanghui Zhao, Jiaming Wang, Yuqi Li, WenBo Xiong, Yifei Chen, Xian Hu, Wei Deng, Daiguo Zhou, Sujith Roy V, Claudia Jesuraj, Vikas B, Spoorthi LC, Nikhil Akalwadi, Ramesh Ashok Tabib, Uma Mudenagudi, Yuxuan Jiang, Chengxi Zeng, Tianhao Peng, Fan Zhang, David Bull Wei Zhou, Linfeng Li, Hongyu Huang, Hoyoung Lee, SangYun Oh, ChangYoung Jeong, Axi Niu, Jinyang Zhang, Zhenguo Wu, Senyan Qing, Jinqiu Sun, Yanning Zhang Abstract This paper provides a review of the NTIRE 2026 challenge on real-world face restoration, highlighting the proposed solutions and the resulting outcomes. The challenge focuses on generating natural and 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. Performance is evaluated using a weighted image quality assessment (IQA) score and employs the AdaFace model as an identity checker. The competition attracted 96 registrants, with 10 teams submitting valid models; ultimately, 9 teams achieved valid scores 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. 1 Introduction Face restoration aims to reconstruct high-quality (HQ) face images from low-quality (LQ) inputs degraded by blur, noise, compression, and other distortions. Since severe degradation often removes a large amount of visual information, this task is inherently ill-posed. Meanwhile, with the continuous progress of portrait imaging technology, users increasingly expect restored face images to exhibit both rich details and high fidelity. This makes it essential for restoration methods to produce outputs that are not only clear but also natural and realistic. In recent years, deep learning has substantially advanced face restoration. Methods based on CNNs, Transformers 98 ; 78 ; 82 ; 66 , GANs 6 ; 74 ; 86 ; 5 , and diffusion models 77 ; 43 ; 84 ; 8 ; 54 ; 61 ; 80 ; 38 ; 91 ; 65 ; 72 demonstrated strong performance. A key challenge in this field lies in how to effectively model face priors. Traditional image restoration methods often rely on statistical priors, whereas modern neural methods tend to learn such priors directly from data. Among them, geometric-prior-based approaches 90 ; 9 ; 29 ; 60 are particularly valuable because they provide explicit structural cues for facial reconstruction. However, when the degradation is relatively mild, users often expect the restored results to remain highly realistic, including subtle skin textures that are usually captured only by high-end imaging devices. Therefore, beyond semantic guidance, texture priors are also vital for recovering fine facial details. Recent studies 16 ; 98 ; 78 ; 82 ; 66 have extensively explored Transformer-based designs for incorporating face priors. Representative methods such as CodeFormer 98 and DAEFR 66 employ codebooks learned from HQ face images as priors. Although these methods are effective at preserving facial information, they still show limitations when handling severely degraded images, especially in transition regions between the face and the background. For more severely degraded inputs, generative capability becomes increasingly important. GAN-based methods 6 ; 74 ; 86 ; 5 have shown strong ability in synthesizing plausible facial details. Among them, GFPGAN 74 is particularly notable, not only for its effective restoration framework, but also for providing benchmark datasets widely used by the computer vision community. More recently, diffusion-based methods 77 ; 43 ; 84 ; 8 ; 54 ; 61 ; 80 ; 38 ; 91 ; 65 ; 72 have emerged as a powerful paradigm. Benefiting from the strong generative priors of diffusion models, high-quality face restoration from severely degraded inputs has become increasingly feasible. DR2 77 transforms the input into noisy states and progressively denoises it to recover essential semantic information. DiffBIR 38 further improves facial detail restoration by leveraging a pre-trained latent diffusion model as a strong prior. In addition, super-resolution models such as SUPIR 89 and StableSR 70 have also been widely adopted in this competition, further highlighting the effectiveness of diffusion-based techniques for real-world face restoration. Very recently, researchers have made significant progress in advancing the field of face restoration. FaceMe 41 and RefSTAR 87 combine reference images with diffusion models, greatly improving reference-based face restoration. InterLCM 32 introduces latent consistency models to the field. It uses a 4-step LCM to improve inference efficiency. OSDFace 72 uses pre-trained models to reduce multi-step diffusion sampling to a single step. This achieves faster inference while maintaining high restoration quality. ( 53 ) uses the Schrödinger Bridge and Pseudo-Hashing to explore optimal transport paths during face restoration. FLIPNET 44 integrates restoration and degradation modes, which offer a new paradigm for learning real-world degradation. SSDiff 33 focuses on old photo restoration. It proposes a training-free method that uses a staged, region-specific guidance scheme. In collaboration with the 2026 New Trends in Image Restoration and Enhancement (NTIRE 2026) workshop, we organized a challenge on real-world face restoration. The challenge aims to recover high-quality face images from degraded low-quality inputs, with an emphasis on richer textures, more realistic facial appearances, and consistent identity preservation. Its goal is to encourage the development of solutions that achieve strong restoration quality with the best perceptual performance, while also revealing current trends in face restoration design. This challenge is one of the challenges associated with the NTIRE 2026 Workshop 1 1 1 https://www.cvlai.net/ntire/2026/ on: deepfake detection 20 , high-resolution depth 92 , multi-exposure image fusion 55 , AI flash portrait 18 , professional image quality assessment 51 , light field super-resolution 76 , 3D content super-resolution 73 , bitstream-corrupted video restoration 99 , X-AIGC quality assessment 42 , shadow removal 68 , ambient lighting normalization 67 , controllable Bokeh rendering 59 , rip current detection and segmentation 14 , low light image enhancement 11 , high FPS video frame interpolation 12 , Night-time dehazing 1 ; 2 , learned ISP with unpaired data 49 , short-form UGC video restoration 34 , raindrop removal for dual-focused images 35 , image super-resolution (x4) 10 , photography retouching transfer 15 , mobile real-word super-resolution 31 , remote sensing infrared super-resolution 39 , AI-Generated image detection 19 , cross-domain few-shot object detection 52 , financial receipt restoration and reasoning 17 , real-world face restoration 71 , reflection removal 4 , anomaly detection of face enhancement 97 , video saliency prediction 45 , efficient super-resolution 56 , 3d restoration and reconstruction in adverse conditions 40 , image denoising 62 , blind computational aberration correction 64 , event-based image deblurring 63 , efficient burst HDR and restoration 48 , low-light enhancement: ‘twilight cowboy’ 28 , and efficient low light image enhancement 83 . 2 NTIRE 2026 Real-world Face Restoration This challenge focuses on restoring real-world degraded face images. The task is to recover high-quality face images with rich high-frequency details from low-quality inputs. At the same time, the output should preserve facial identity to a reasonable degree. There are

Read the complete text