This paper presents NTIRE 2026, the 3rd Restore Any Image Model (RAIM) challenge on multi-exposure image fusion in dynamic scenes. We introduce a benchmark that targets a practical yet difficult HDR imaging setting, where exposure bracketing must be fused under scene motion, illumination variation, and handheld camera jitter. The challenge data contains 100 training sequences with 7 exposure levels and 100 test sequences with 5 exposure levels, reflecting real-world scenarios that frequently cause misalignment and ghosting artefacts. We evaluate submissions with a leaderboard score derived from PSNR, SSIM, and LPIPS, while also considering perceptual quality, efficiency, and reproducibility during the final review. This track attracted 114 participating teams and received 987 submissions. The winning methods significantly improved the ability to remove artifacts from multi-exposure fusion and recover fine details. The dataset and the code of each team can be found at the repository: https://github.com/qulishen/RAIM-HDR.
Figures & tables
Figure 1: Representative image sequences in RAIM MEF in dynamic scenes. The data includes various scenes and times, with a particular focus on scenes with high dynamic range.
Figure 2: Overview of the WHU-VIP method with the AFUNet backbone, the SFT module, and the DCM-RG module.
Figure 3: Overview of the SHL team’s proposed framework. The method introduces a lightweight global guidance branch that provides context from the downsampled full-resolution image, complemented by a Neighboring Pixel Relationships (NPR) loss to preserve high-frequency local details.
Figure 4: Overview of the nunucccb method.
Figure 5: Overview of the untrafusion pipeline.
Team Name
Test Stage 1
Test Stage 2
Final Score ↑
PSNR ↑
SSIM ↑
LPIPS ↓
Score ↑
PSNR ↑
SSIM ↑
LPIPS ↓
Score ↑
WHU-VIP
27.059
0.915
0.089
58.249
27.556
0.923
0.080
59.529
58.889
SHL
27.653
0.920
0.112
57.110
28.825
0.928
0.101
58.958
58.034
nunucccb
26.839
0.913
0.111
56.395
27.011
0.918
0.099
57.605
57.000
untrafusion
26.847
0.908
0.095
57.333
26.002
0.907
0.122
54.733
56.033
I^2 Group & Transsion
26.221
0.904
0.121
54.853
26.137
0.909
0.113
55.619
55.236
Table 1: Verified final results for RAIM Track 2. The ranking is computed from validated scores on Test Stage 1 and Test Stage 2.
Figure 6: Overview of the I2 Group & Transsion method.
Figure 7: TimeDiffiT architecture [ 51 ] . The doubly-corrupted input X~tM and noise level σt enter the time-conditioned U-Net (encoder: orange; decoder: blue), producing the restored output X^ .
Figure 8: Visual comparison on representative examples from Test Stage 1. The compared methods show clear differences in structural consistency, detail preservation, and ghosting suppression.
Figure 9: Visual comparison on representative examples from Test Stage 2. This stage further highlights differences in motion handling, exposure robustness, and artifact suppression.
Multi-exposure is an efficient way to capture real-world high-dynamic-range (HDR) scenes. However, HDR imaging suffers from severe ghosting artifacts in dynamic scenes due to the temporal gap between sequential exposures. In this article, we categorize the literature on two important topics on HDR imaging: multi-exposure fusion (MEF) and ghost removal. Conventional filter-based and data-driven methods are studied in pixel space and feature space. For popular deep learning-based approaches, we provide a granular taxonomy based on their alignment and fusion domains: pixel-space methods, which typically employ explicit motion compensation such as optical flow or spatial transformers, and feature-space methods, which leverage implicit alignment through deformable convolutions, attention mechanisms, or latent representation merging. Representative works are compared across different supervision settings, and key design principles are summarized. In addition, this survey summarizes commonly used datasets and evaluation metrics, discussing their applicability under diverse output forms. Finally, major bottlenecks and promising directions for future research are outlined.
Qian Tao, Wei Wang, Chaobing Zheng +1
School of Computer Science and Technology, Hubei Province Key Laboratory of Intelligent Information Processing and Real-Time Industrial System, Wuhan University of Science and Technology, Wuhan 430081, China · School of Electronic Information, Wuhan University of Science and Technology, Wuhan 430081, China · Institute of Advanced Intelligence and Computing, A*STAR, Singapore 138632
In this paper, we present a comprehensive overview of the NTIRE 2026 3rd Restore Any Image Model (RAIM) challenge, with a specific focus on Track 3: AI Flash Portrait. Despite significant advancements in deep learning for image restoration, existing models still encounter substantial challenges in real-world low-light portrait scenarios. Specifically, they struggle to achieve an optimal balance among noise suppression, detail preservation, and faithful illumination and color reproduction. To bridge this gap, this challenge aims to establish a novel benchmark for real-world low-light portrait restoration. We comprehensively evaluate the proposed algorithms utilizing a hybrid evaluation system that integrates objective quantitative metrics with rigorous subjective assessment protocols. For this competition, we provide a dataset containing 800 groups of real-captured low-light portrait data. Each group consists of a 1K-resolution low-light input image, a 1K ground truth (GT), and a 1K person mask. This challenge has garnered widespread attention from both academia and industry, attracting over 100 participating teams and receiving more than 3,000 valid submissions. This report details the motivation behind the challenge, the dataset construction process, the evaluation metrics, and the various phases of the competition. The released dataset and baseline code for this track are publicly available from the same GitHub repository, and the official challenge webpage is hosted on CodaBench.
Multi-Exposure Fusion (MEF) effectively extends dynamic range, but practical deployment is hindered by motion-induced ghosting and the scarcity of high-quality dynamic benchmarks. Current benchmarks largely neglect dynamic scenes and lack reliable ground truth, making it difficult to handle the complexity of real-world motions. In response, we introduce ExpoMotion, a large-scale benchmark designed to evaluate deghosting capabilities. Comprising 1,738 sequences and 10,909 images across diverse environments, it covers a wide range of motions and provides high-fidelity GTs constructed through an expert-guided acquisition pipeline. To tackle the complex dynamics and extreme conditions captured in this benchmark, we propose the Householder Orthogonal Projection network (HOP), which revisits MEF deghosting from a mathematical perspective via Householder transformation, decoupling multi-frame alignment into exposure pre-alignment and ghost filtering. Specifically, the Global Priors Illumination Alignment (GPIA) module first rectifies drastic dynamic range discrepancies by utilizing global statistics for exposure harmonization. Regarding ghost removal, our Householder Orthogonal Attention (HOA) models artifacts as orthogonal perturbations. By employing a dynamic Householder reflector, HOA effectively projects ghosts out of the feature manifold while preserving high-frequency details. Experiments demonstrate that our ExpoMotion dataset enables superior generalization and artifact-free detail restoration, while also validating the effectiveness and efficiency of the HOP method. The dataset and code are available at https://github.com/Leo-LiuYao/ExpoMotion.
Yao Liu, Lishen Qu, Shihao Zhou +6
Nankai International Advanced Research Institute (SHENZHEN·FUTIAN) · College of Computer Science, Nankai University · OPPO Research Institute +2