RAW-Flow: Advancing RGB-to-RAW Image Reconstruction with Deterministic Latent Flow Matching

TL;DR

RAW-Flow introduces a novel deterministic latent flow matching framework for RGB-to-RAW image reconstruction, effectively addressing detail and color fidelity issues by modeling the task as a latent transport problem.

Contribution

The paper pioneers a generative approach using flow matching in latent space, incorporating cross-scale guidance and a dual-domain autoencoder for improved RAW reconstruction.

Findings

Outperforms state-of-the-art methods quantitatively.

Achieves superior visual quality in reconstructed RAW images.

Effectively preserves structural details and color fidelity.

Abstract

RGB-to-RAW reconstruction, or the reverse modeling of a camera Image Signal Processing (ISP) pipeline, aims to recover high-fidelity RAW data from RGB images. Despite notable progress, existing learning-based methods typically treat this task as a direct regression objective and struggle with detail inconsistency and color deviation, due to the ill-posed nature of inverse ISP and the inherent information loss in quantized RGB images. To address these limitations, we pioneer a generative perspective by reformulating RGB-to-RAW reconstruction as a deterministic latent transport problem and introduce a novel framework named RAW-Flow, which leverages flow matching to learn a deterministic vector field in latent space, to effectively bridge the gap between RGB and RAW representations and enable accurate reconstruction of structural details and color information. To further enhance latent transport, we introduce a cross-scale context guidance module that injects hierarchical RGB features into the flow estimation process. Moreover, we design a dual-domain latent autoencoder with a feature alignment constraint to support the proposed latent transport framework, which jointly encodes RGB and RAW inputs while promoting stable training and high-fidelity reconstruction. Extensive experiments demonstrate that RAW-Flow outperforms state-of-the-art approaches both quantitatively and visually.