This project is under active development, please stay tuned! ☕

Sep 29, 2024: Code & models of ProRes are released, and some minor issues have been fixed. We will update the arXiv version with more details upon acceptance.

Aug 28, 2024: Our newest paper Perceive-IR is on arXiv and under peer review! Perceive-IR performs well under two all-in-one settings.

Jun 26, 2023: We've released ProRes! We wish this work would inspire more works on prompt learning with image restoration tasks. Code & models are coming soon!

• ProRes addresses universal image restoration with degradation-aware prompts, which is the first prompt-based versatile framework for image restoration.

• ProRes demonstrate two superior capabilities: (1) control ability for desired outputs and (2) transferability based on prompt tuning.

• ProRes can be easily adapted for new tasks or new datasets through effective and efficient prompt tuning.

• Specific prompts can control the output of ProRes. Moreover, combining different prompts can tackle the images with multiple corruptions.

• Linux, CUDA>=9.2, GCC>=5.4
• PyTorch >= 1.8.1
• MATLAB for evaluation

Install detectron2, following the instructions in here. Or simply use the following command.

cd ProRes/
git clone https://github.com/facebookresearch/detectron2
python -m pip install -e detectron2

• Download the denoising dataset from SIDD.
• Download the low-light enhancement dataset from LoL.
• Download the deraining dataset from Synthetic Rain Datasets.
• Download the deblurring dataset from Synthetic Blur Datasets.

Run the following commands to generate corresponding JSON files for each dataset.

#denoising
python data/sidd/gen_json_sidd.py --split train
python data/sidd/gen_json_sidd.py --split val

# low-light enhancement
python data/lol/gen_json_lol.py --split train
python data/lol/gen_json_lol.py --split val

# deraining
python data/derain/gen_json_rain.py --split train
python data/derain/gen_json_rain.py --split val

# derblurring
python data/derain/gen_json_blur.py --split train
python data/derain/gen_json_blur.py --split val

We recommend the dataset directory structure to be the following:

$ProRes/datasets/
    denoise/
        train/
        val/
    enhance/
        our485/
            low/
            high/
        eval15/
            low/
            high/
    derain/
        train/
            input/
            target/
        test/
            Rain100H/
            Rain100L/
            Test100/
            Test1200/
            Test2800/
    deblur/
        train/
            input/
            target/
        test/
            GoPro/
            HIDE/
            RealBlur_J/
            RealBlur_R/

    target-derain_train.json
    gt-enhance_lol_train.json
    groundtruth-denoise_ssid_train448.json
    groundtruth_crop-deblur_gopro_train.json
    
    target-derain_test_rain100h.json
    gt-enhance_lol_eval.json
    groundtruth-denoise_ssid_val256.json
    groundtruth-deblur_gopro_val.json

Download pre-trained MAE ViT-Large model and put it in the folder ./ProRes/pretrained_weights/.

sh scripts/train.sh

Download the pretrained model and put it in the folder ./ProRes/models/prores_vitl_pretrained_sl1_mprnetprompt_add/.

Run the following command:

sh eval_ours.sh

Notes:

• The works we used for reference including Uformer(paper,code), MPRNet(paper,code), MIRNet-v2(paper,code), Restormer(paper,code), MAXIM(paper,code) and Painter(paper,code).
• For both Painter and ProRes, we adopt ViT-Large with MAE pre-trained weights.
• More experimental results are listed in the paper!

Visualization results processed from images of different corruptions. Compared with the original inputs, the outputs are consistent with the given visual prompts.

Visualization results processed by different prompts. Compared with the original inputs, the outputs remain unchanged with irrelevant visual prompts.

Visualization results processed by ProRes from images of mixed types of degradation, i.e., low-light and rainy. ProRes adopts two visual prompts for low-light enhancement (E) and deraining (D) and combines the two visual prompts by linear weighted sum, i.e., αD (1 − α)E, to control the restoration process.

Visualization results of ProRes on the FiveK dataset. We adopt two settings, i.e., direct inference and prompt tuning, to evaluate ProRes on the FiveK dataset (a new dataset for low-light enhancement).

Visualization results of ProRes on the RESIDE-6K dataset via prompt tuning for image dehazing (a new task).

If you find our paper and code useful for your research, please consider giving a star ⭐ and citation 📝 :

@article{
      title={ProRes: Exploring Degradation-aware Visual Prompt for Universal Image Restoration}, 
      author={Jiaqi Ma and Tianheng Cheng and Guoli Wang and Xinggang Wang and Qian Zhang and Lefei Zhang},
      journal={arXiv preprint arXiv:2306.13653},
      year={2023}
}

This project is based on MAE, BEiT, MIRNet, MPRNet, Uformer and Painter. Thanks for their wonderful work!