← Search

Diedong Feng

2 accepted papers

2026

Bridging RGB and RAW: Single-step Deterministic Flow with Homogeneous Aligned Guidance

ICML 2026poster

Reconstructing high-fidelity RAW sensor data from processed RGB images is a fundamental yet ill-posed problem, plagued by irreversible information loss and complex non-linear ISP transformations. While generative models offer high-quality reconstruction, they suffer from prohibitive computational co…

Cited by 0SourceScholar
2026

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

AAAI 2026technical

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 still struggle with

Cited by 0SourcePDFScholar