/facial-image-super-resolution

Primary LanguagePythonMIT LicenseMIT

image-super-resolution

This project uses the MMEditing library from Open-MMLab to improve the resolution of facial images.

https://github.com/open-mmlab

MODULES

Module for generating HQ test images from the SRResNet model.
test.py

Module for generating the training annotation file.
annotation.py

Module for reading json logs files and plotting the training curves in the report.
plot.py

Modifed configuration file for SRResNet.
srresnet_project.py

Modified configuration file for RRDBNet.
esrgan_project_psnr.py

Modified configuration file for ESRGAN module (for future work).
esrgan_project.py

FILES

Best SRResNet model.
./best_model/iter_285.pth

Annotation file.
./data/train_ann.txt

Training logs (txt and json) for SRResNet.
./srresnet_logs/*

Complete training logs used to plot the training curves for SRResNet.
./srresnet_logs/srresnet.json

Training logs (txt and json) for RRDBNet.
./rrdbnet_logs/*

Complete training logs used to plot the training curves for RRDBNet.
./rrdbnet_logs/rrdbnet.json

Links to Checkpoint File and HQ images:

https://drive.google.com/file/d/1Ffcq7V21e2C4iK5Kb1q1Qi_TSaHVTiTV/view?usp=sharing

https://drive.google.com/file/d/1cUQm1J2kTorAcsqoXzRvMB75HSZrKVrn/view?usp=sharing

SETUP

cd image_super_resolution/project

Create and activate a conda virtual environment.
conda create --name cv python=3.8 -y
conda activate cv

Install the GPU version of PyTorch and Torchvision.
conda install pytorch=1.11.0 torchvision cudatoolkit=10.2 -c pytorch

Install MMCV for GPU.
pip install mmcv-full -f https://download.openmmlab.com/mmcv/dist/cu102/torch1.11.0/index.html

Install MMEditing.
git clone https://github.com/open-mmlab/mmediting.git
cd mmediting
pip install -v -e

Go back to the current directory.
cd ..

Move all directories and files in this package to the current directory. The directory structure will be:

./best_model/
	iter_285.pth
./data/
	test/
		HQ/
		LQ/
	train/
		GT/
		LQ/
	val/
		GT/
		LQ/
	train_ann.txt
./rrdbnet_logs/
	rrdbnet.json
	*
./srresnet_logs/
	srresnet.json
	*
./mmediting/
	*
esrgan_project.py
esrgan_project_psnr.py
srresnet_project.py
annotation.py
plot.py
test.py