There and Back Again: A Flexible-Frame Transformer for Multi-Exposure Fusion

2026-06-26

AI Summary

This work addresses the limitation of existing multi-exposure fusion methods, which typically support only a fixed number of input frames and thus lack flexibility in real-world scenarios involving variable numbers of exposures. To overcome this, we propose FreeMEF—the first Transformer-based multi-exposure fusion model capable of processing an arbitrary number of input frames without requiring retraining or architectural modifications. FreeMEF introduces three key innovations: a Recurrent State Space Module (RSSM), Extremum-Aware Hybrid Attention (EAHA), and an Affine-injected Feed-Forward Network (AFFN), collectively mitigating the similarity paradox and enhancing luminance and contrast fidelity. Extensive experiments demonstrate that FreeMEF consistently outperforms state-of-the-art methods across three benchmark datasets, with both quantitative metrics and qualitative assessments confirming its superior fusion performance.

Abstract

Multi-exposure fusion (MEF) brings the dynamic range of conventional cameras closer to that of human vision, producing images with rich scene content. Given the large variability in scene luminance, exposure strategies often require different numbers of frames to capture the full radiance range faithfully. However, conventional MEF techniques are typically designed for a fixed number of inputs, forcing deployment systems to maintain separate models for different frame-count requirements, which undermines deployment efficiency. To address this limitation, we propose FreeMEF, the first flexible-frame transformer for MEF that seamlessly accommodates varying numbers of input exposures without retraining or architectural changes. The proposed approach consists of two key modules. First, we introduce a recurrent state space module (RSSM) that sequentially fuses features from arbitrary sequences via adaptive alignment and state-space recurrent modeling, thereby providing global information guidance for the subsequent restoration. Second, we devise a global feature guided block (GFGB) incorporating an extremity-aware hybrid attention (EAHA) and an affine-injection feed-forward network (AFFN), which effectively resolves the similarity paradox while simultaneously optimizing contrast and brightness regulation. Extensive experiments on three benchmark datasets demonstrate the effectiveness of our method, which performs favorably against state-of-the-art methods both quantitatively and qualitatively.

Problem

• multi-exposure fusion
• flexible-frame input
• dynamic range
• deployment efficiency
• variable exposure count

Innovation

• Flexible-frame Transformer
• Multi-exposure Fusion
• Recurrent State Space Module
• Global Feature Guided Block
• Extremity-aware Hybrid Attention