RIRNet

Source codes for paper "RIRNet: Recurrent-In-Recurrent Network for Video Quality Assessment" in Proceedings of the 28th ACM International Conference on Multimedia (ACM MM ’20).

Usages

Testing a single sample

Predicting video quality with our model trained on the KoNViD-1k Dataset (coming later).

python ./released/demo.py

You will get a quality score ranging from 0-5, and a higher value indicates better percerptual quality.

Training on VQA databases

Reading mos values from the .csv files:

python ./Released/get_label.py

Processing .mp4 files to frames:

python ./Released/get_frame.py

Training the model with 'label_path' and 'frame_path':

python ./Released/source.py

Environment

• Python 3.6.5
• Pytorch 1.0.1
• Cuda 9.0 Cudnn 7.1

Citation

If you find this work useful for your research, please cite our paper:

@inproceedings{chen2020RIRNet,
  title={RIRNet: Recurrent-in-Recurrent Network for Video Quality Assessment},
  author={Chen, Pengfei and Li, Leida and Ma, Lei and Wu, Jinjian and Shi, Guangming},
  booktitle={Proceedings of the 28th ACM International Conference on Multimedia},
  pages={834--842},
  year={2020}
}