ProDehaze: Prompting Diffusion Models Toward Faithful Image Dehazing
Preprint
Abstract
Recent approaches using large-scale pretrained diffusion models for image dehazing improve perceptual quality but often suffer from hallucination issues, producing unfaithful dehazed image to the original one. To mitigate this, we propose ProDehaze, a framework that employs internal image priors to direct external priors encoded in pretrained models. We introduce two types of \textit{selective} internal priors that prompt the model to concentrate on critical image areas: a Structure-Prompted Restorer in the latent space that emphasizes structure-rich regions, and a Haze-Aware Self-Correcting Refiner in the decoding process to align distributions between clearer input regions and the output. Extensive experiments on real-world datasets demonstrate that ProDehaze achieves high-fidelity results in image dehazing, particularly in reducing color shifts. Our code is at https://github.com/TianwenZhou/ProDehaze.Keywords
All Related Versions
This publication has 20 references indexed in Scilit:
- DiffBIR: Toward Blind Image Restoration with Generative Diffusion Prior2024
- LID-Net: A lightweight image dehazing network for automatic driving vision systemsDigital Signal Processing, 2024
- Exploiting Diffusion Prior for Real-World Image Super-ResolutionInternational Journal of Computer Vision, 2024
- Compensation Atmospheric Scattering Model and Two-Branch Network for Single Image DehazingIEEE Transactions on Emerging Topics in Computational Intelligence, 2024
- Frequency compensated diffusion model for real-scene dehazingNeural Networks, 2024
- Frequency and content dual stream network for image dehazingImage and Vision Computing, 2023
- Taming Transformers for High-Resolution Image Synthesis2021
- NIMA: Neural Image AssessmentIEEE Transactions on Image Processing, 2018
- The CIEDE2000 color-difference formula: Implementation notes, supplementary test data, and mathematical observationsColor Research & Application, 2004
- A theory for multiresolution signal decomposition: the wavelet representationIEEE Transactions on Pattern Analysis and Machine Intelligence, 1989