search

Zhaozixiang1228 / MMIF-DDFM

[ICCV 2023 Oral] Official implementation for "DDFM: Denoising Diffusion Model for Multi-Modality Image Fusion."
207 stars 21 forks source link
denoising-diffusion generative-models image-fusion low-level-vision

readme

DDFM (ICCV 2023 Oral)

Codes for DDFM: Denoising Diffusion Model for Multi-Modality Image Fusion. (ICCV 2023 Oral)

Zixiang Zhao, [Haowen Bai](), Yuanzhi Zhu, Jiangshe Zhang, Shuang Xu, Yulun Zhang, Kai Zhang, Deyu Meng, Radu Timofte and Luc Van Gool.

-[Paper]
-[ArXiv]
-[Supplementary Materials]

Update

Citation

@InProceedings{Zhao_2023_ICCV,
    author    = {Zhao, Zixiang and Bai, Haowen and Zhu, Yuanzhi and Zhang, Jiangshe and Xu, Shuang and Zhang, Yulun and Zhang, Kai and Meng, Deyu and Timofte, Radu and Van Gool, Luc},
    title     = {DDFM: Denoising Diffusion Model for Multi-Modality Image Fusion},
    booktitle = {Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)},
    month     = {October},
    year      = {2023},
    pages     = {8082-8093}
}

Abstract

Multi-modality image fusion aims to combine different modalities to produce fused images that retain the complementary features of each modality, such as functional highlights and texture details. To leverage strong generative priors and address challenges such as unstable training and lack of interpretability for GAN-based generative methods, we propose a novel fusion algorithm based on the denoising diffusion probabilistic model (DDPM). The fusion task is formulated as a conditional generation problem under the DDPM sampling framework, which is further divided into an unconditional generation subproblem and a maximum likelihood subproblem. The latter is modeled in a hierarchical Bayesian manner with latent variables and inferred by the expectation-maximization (EM) algorithm. By integrating the inference solution into the diffusion sampling iteration, our method can generate high-quality fused images with natural image generative priors and cross-modality information from source images. Note that all we required is an unconditional pre-trained generative model, and no fine-tuning is needed. Our extensive experiments indicate that our approach yields promising fusion results in infrared-visible image fusion and medical image fusion.

🌐 Usage

⚙ 1. Virtual Environment

# create virtual environment
conda create -n DDFM python=3.8.10
conda activate DDFM
# select pytorch version yourself
# install DDFM requirements
pip install -r requirements.txt

📃 2. Pre-trained Checkpoint Preparation

From the link, download the checkpoint "256x256_diffusion_uncond.pt" and paste it to './models/'.

🏊 3. Data Preparation

Download the Infrared-Visible Fusion (IVF) and Medical Image Fusion (MIF) dataset and place the paired images in the folder './input/'.

🏄 4. Inference (Sampling)

If you want to infer with our DDFM and obtain the fusion results in our paper, please run

python sample.py

Then, the fused results will be saved in the './output/recon/' folder.

Additionally,

🙌 DDFM

Illustration of our DDFM model.

Detail of DDFM.

Qualitative fusion results.

Quantitative fusion results.

Infrared-Visible Image Fusion

Medical Image Fusion

📖 Related Work