Unofficial Tensorflow implementation of "Image Super-Resolution Using Very Deep Residual Channel Attention Networks", ECCV 2018, [arXiv]
-
Training Dataset is 800 training + 100 validtion images of DIV2K dataset.The TFRecords for the same and npz files of the benchmark dataset can be found in the drive link.
-
Maintain the following Directory structure
-
cd to
code
. Specify the scale of upsampling by '--scale'. Default value is 4. -
Run
python3 training.py
Run python3 infer.py
. The output will be stored in results
.
-
cd to eval
-
Run
./matlab_eval.sh
Note that all the results shown below are obtained from limited period of training. Accuracy can be improved when trained for more duration.
Ground Truth | Predicted |
---|---|
Results obtained are in Y channel. We can perform Bicubic interpolation in Cr and Cb channels and merge with the predicted Y channel.