Super resolution in the compressed domain of learning-based image codecs

CDSR stands for Compressed Domain Super Resolution.

Setting up

Docker

Build a Docker container:

sudo docker build . -t cdsr

Run the container:

sudo docker run --ipc=host --gpus all --rm -it \
    -v ${PWD}:/workspace \
    -w /workspace \
    cdsr bash

Alternatively run a Docker container for development:

sudo docker run --ipc=host --gpus all --rm -it \
    -p 0.0.0.0:6006:6006 -p 0.0.0.0:8888:8888 \
    --hostname cdsr-dev --name cdsr-dev \
    -v ${PWD}:/workspace \
    -w /workspace \
    cdsr bash

Testing

The pretrained models and test dataset can be downloaded from the following FTP by using dedicated FTP clients, such as FileZilla or FireFTP (we recommend to use FileZilla):

Protocol: FTP
FTP address: tremplin.epfl.ch
Username: datasets@mmspgdata.epfl.ch
Password: ohsh9jah4T
FTP port: 21

Download pretrained models pretrained_model.tgz from the cdsr folder in the FTP server. Extract files:

tar -xzf pretrained_model.tgz

Download the test dataset jpegai.tgz from the cdsr folder in the FTP server. Extract the files from jpegai.tgz to the datasets folder:

mkdir datasets
cd datasets
tar -xzf jpegai.tgz

Inside the Docker container, run the following command:

cd src
python test.py -opt options/jpegai/test_jpegai.yml

The upscaled images will be found in the results folder.