Single Image Super Resolution

This repo contains the implementations of some common models used for SISR.

For simple inference tasks, download the pretrained models from here. Place them in the pretrained/ directory.

Next, configure the yaml config file in configs/ directory for the respective model you wish to run and then run inference as:

python main.py --config-file <config file>

For example:

python main.py --config-file RCAN.yaml

Report

A complete summary report of these models can be found here: report_link

References:

• Enhanced Deep Super Resolution Network (EDSR) : link
• Residual Channel Attention Network (RCAN) : link

Note: The code for GAN based models is not currently released. It will be released later.