-
Code release - Upload datasets
log
: save training lognets
: define iqa modelsave
: save modelnets
: define iqa modeldatasets.py
: define datasetsdatasets_deepsrq.py
: define datasets for deepsrqengine.py
: training and test enginetrain_test_IQA.py
: setup training and test
python train_test_IQA.py --dataset xxx --netFile xxx --gpuid x --batch_size 64
Some mandatory options:
--dataset
: string, Training and testing dataset, support datasets: 'CVIU' | 'QADS'| 'Waterloo'.--netFile
: string, IQA model, support models: 'DBCNN' | 'HyperIQA' | 'CNNIQA' | 'Resnet50' | 'JCSAN' | 'DeepSRQ'.--gpuid
: int, gpu device--batch_size
: int, Batch size, 64.
This project is based on HyperIQA. Thanks for the awesome work.
Please cite the following paper if you use this repository in your reseach.
@article{fu2023scale,
title={Scale Guided Hypernetwork for Blind Super-Resolution Image Quality Assessment},
author={Fu, Jun},
journal={arXiv preprint arXiv:2306.02398},
year={2023}
}
For any questions, feel free to contact: fujun@mail.ustc.edu.cn