/TISR_APTX4869

Source code of team APTX4869 (3rd in x2 MR2HR PSNR/SSIM metrics) in Thermal Image Super-Resolution Challenge Results - PBVS 2023

Primary LanguagePython

TISR_APTX4869

Source code of team APTX4869 (3rd x2 MR2HR) in Thermal Image Super-Resolution Challenge Results - PBVS 2023

This codebase is based on BasicSR. Thanks for the effort and great work of the authors.

Preprocess

change the paths in process_scripts/alignment_tisr_mr2hr.py and run the script to align the MR and HR thermal images first.

Run train and test

change the paths in the options_train/exp_name.yml and options_test/exp_name.yml to your own dataset location, and then run the bash scripts to train the model.

sh run_train_track1.sh

after the model is trained, use the following command to conduct inference

python test_track1.py -opt options_test/exp_name.yml

you can also use the trained result in GoogleDrive to inference the testset directly.