LLaVA-based semantic feature modulation diffusion model for underwater image enhancement

Open Webpage

Guodong Fan, Shengning Zhou, Zhen Hua, Jinjiang Li, Jingchun Zhou

Published: 01 Jan 2026, Last Modified: 09 Nov 2025Inf. Fusion 2026EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•We propose LSFM-Diff, a diffusion model for UIE with dual LLaVA semantic guidance.•We introduce WTIF-CR, a module that fuses text and features for fine-grained guidance.•We design SGDA, a mechanism for spatially adaptive feature enhancement within UNet.