/DMT-Net

Image dehazing

Primary LanguagePython

From Synthetic to Real: Image Dehazing Collaborating with Unlabeled Real Data

Official implementation of From Synthetic to Real: Image Dehazing Collaborating with Unlabeled Real Data(ACM MM 2021) arXiv link

Citation

@inproceedings{liu21DMT-Net,
    author = {Liu, Ye and Zhu, Lei and Pei, Shunda and Fu, Huazhu and Qin, Jing and Zhang, Qing and Wan, Liang and Feng, Wei},
    title = {From Synthetic to Real: Image Dehazing Collaborating with Unlabeled Real Data},
    booktitle = {ACM MM},
    year = {2021}
}

Haze4K dataset

Haze4K dataset is available at Haze4K link(pw: cmmr)

Results in SOTS, Haze4K and HazeRD datasets

Quantitative Results

Visual Results on synthetic dataset Haze4K

Visual Results on real world images