NTIRE 2025 Challenge on Efficient Burst HDR and Restoration: Datasets, Methods, and Results

2025-05-17

AI Summary

This work addresses lightweight multi-frame HDR synthesis and image restoration by introducing the first burst HDR challenge under strict efficiency constraints (parameters <30M, computational cost <4.0T FLOPs). To enable fair evaluation of efficient algorithms, we construct the first dedicated multi-frame RAW benchmark dataset, comprising nine unaligned, noisy, multi-exposure RAW images per scene. Methodologically, our framework integrates differentiable multi-frame registration, RAW-domain noise modeling, lightweight CNN/Transformer architectures, exposure-aware feature fusion, and end-to-end HDR reconstruction. The challenge attracted 217 participating teams, with six submitting valid solutions. The winning method achieved a PSNR of 43.22 dB, demonstrating the feasibility and state-of-the-art performance of low-complexity burst HDR synthesis. This work advances efficient HDR fusion paradigms and establishes a new standard for resource-constrained HDR imaging research.

Technology Category

• Computer Vision: Computational Photography, Image & Video Synthesis
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Security and Privacy: Data transparency and provenance
• Systems and Infrastructure for Web, Mobile and WoT: Web performance, measurement, and characterization

Abstract

This paper reviews the NTIRE 2025 Efficient Burst HDR and Restoration Challenge, which aims to advance efficient multi-frame high dynamic range (HDR) and restoration techniques. The challenge is based on a novel RAW multi-frame fusion dataset, comprising nine noisy and misaligned RAW frames with various exposure levels per scene. Participants were tasked with developing solutions capable of effectively fusing these frames while adhering to strict efficiency constraints: fewer than 30 million model parameters and a computational budget under 4.0 trillion FLOPs. A total of 217 participants registered, with six teams finally submitting valid solutions. The top-performing approach achieved a PSNR of 43.22 dB, showcasing the potential of novel methods in this domain. This paper provides a comprehensive overview of the challenge, compares the proposed solutions, and serves as a valuable reference for researchers and practitioners in efficient burst HDR and restoration.

Problem

• Advance efficient multi-frame HDR and restoration techniques
• Fuse noisy, misaligned RAW frames with varying exposures
• Meet strict efficiency constraints in model parameters and FLOPs

Innovation

• Novel RAW multi-frame fusion dataset
• Efficient fusion under 30M parameters
• Top solution achieves 43.22 dB PSNR