Multi-Scale Fusion and Decomposition Network for Single Image Deraining (MFDNet)

This is an implementation of the MFDNet model.

Installation

The model is built in Python 3.7, PyTorch 1.13.1, CUDA11.6.

For installing, follow these intructions

conda env create -f mfd.yml

Training

• Download training dataset ((raw images)Baidu Cloud, (Password:4qnh) (.npy)Baidu Cloud, (Password:gd2s)) and place it in ./Datasets/train/

• Train the model with default arguments by running

python train.py

Evaluation

1. The pretrained_model is provided in ./checkpoints/checkpoints_mfd.pth

2. Download test datasets (Test100, Rain100H, Rain100L, Test1200, Test2800) from here and place them in ./Datasets/test/

3. Run

python test.py