Newest papers

GLOW-FM: Efficient Generative Low-Light Image Enhancement with One-Step Latent Flow Matching

Ariel Lapid, Yuval Haitman, Reuven Peretz, Arnon Netzer, Idit Diamant · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.31222 · cited by 0 works (OpenAlex) · open-access PDF

Portuguese

Os métodos gerativos recentes baseados em difusão alcançam uma forte qualidade perceptiva ao aproveitar prefixos de imagem poderosos, mas sua amostragem iterativa incorre em custos computacionais substanciais, limitando a implantação prática. Apesar dos recentes ganhos de eficiência em LLIE gerativo, o aprimoramento de alta qualidade com uma avaliação pública de função única e computação global baixa permanece desafiador, particularmente sob condições de baixa luz do mundo real diversas. Introduzimos GLOW-FM, um framework de correspondência de fluxo latente de um passo para a implantação eficiente de LLIE. A GLOW-FM realiza o transporte condicional no espaço latente de um Tiny-EVA congelado e alavanca a Avaliação de Velocidade Terminal (TVM) para permitir a avaliação específica com

Machine translation (Meta M2M100); the English original follows.

Abstract

Low-light image enhancement (LLIE) aims to recover well-exposed and perceptually faithful images from severely degraded observations. Recent diffusion-based generative methods achieve strong perceptual quality by leveraging powerful image priors, but their iterative sampling incurs substantial computational cost, limiting practical deployment. Despite recent efficiency gains in generative LLIE, high-quality enhancement with a single function evaluation and low overall computation remains challenging, particularly under diverse real-world low-light conditions. We introduce GLOW-FM, a one-step latent flow-matching framework for efficient generative LLIE. GLOW-FM performs conditional transport in the latent space of a frozen Tiny-VAE and leverages Terminal Velocity Matching (TVM) to enable enhancement with a single function evaluation. Matched comparisons with alternative one-step objectives support TVM as an effective design choice, and, to the best of our knowledge, this is its first adaptation to image restoration and specifically to LLIE. Extensive experiments on challenging real-world benchmarks demonstrate that GLOW-FM achieves competitive image quality and state-of-the-art cross-dataset generalization while requiring only a single function evaluation and 34 GFLOPs for the small variant. These findings highlight the potential of GLOW-FM to bring generative LLIE closer to practical deployment through single-step inference and compact latent modeling. Code will be made publicly available upon publication.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Computer science, Inference, Artificial intelligence, Matching (statistics), Generative model, Image (mathematics)

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange