DM2F-Enhancing

Final Dehazing Project of Digital Image Processing (COMP130032): Improvement of DM2F-Net.

Installation

1. Clone our repository

git clone https://github.com/linghuyuhangyuan/DM2F-Enhancing.git
cd DM2F-Enhancing

2. Make conda environment

conda create -n DM2F python=3.8
conda activate DM2F

pip install -r requirements.txt

Data Preparation

• HazeRD
• OHaze

Please put them in the data directory.

Pretrained model

Download pretrained model from Google Drive and place them in the ckpts directory.

Training

python train_alg1.py --gpus 0 \
    --load-dm2f ckpts/ckpt/O-Haze/iter_20000_loss_0.04937_lr_0.000000.pth \
    --ckpt-path ckpts/ckpt_alg1

Before training algorithm 1, assume that pretrained model has been put in the ckpts directory.

python train_alg2.py --gpus 1 \
    --ckpt-path ckpts/ckpt_alg2

Evaluation

Before testing, assume that pretrained model has been put in the ckpts directory.

python test_alg1.py

python test_alg2.py

Contact

If you have any question or suggestion, please contact [email protected].