AVAQ

This is the dataset used in paper Deep Multimodality Learning for UAV Video Aesthetic Quality Assessment.

Introduction

AVAQ6000 is a dataset used for Aerial Video Aesthetic Quality assessment. We collected 6,000 UAV video shots, including 3,000 professional shots and 3,000 amateur shots. The professional shots are collected from several documentaries and films. The amateur shots are downloaded from aerial video websites where amateurs can share their works with others.

Each video is an individual shot. Additionally, to avoid the influence of the shot length, the duration time of each shot does not exceed 1 min. The total length of AVAQ6000 is more than 25 hours. The resolution of videos is 720P or 1080P. The framerate is 30 fps.

Download

百度网盘链接: https://pan.baidu.com/s/1pTimrppklDE5ilXfqqTVKw 提取码: j8t8

Citation

Please consider citing the following paper if the code or dataset is helpful in your research work:

@article{kuang2019deep,
  title={Deep Multimodality Learning for UAV Video Aesthetic Quality Assessment},
  author={Kuang, Qi and Jin, Xin and Zhao, Qinping and Zhou, Bin},
  journal={IEEE Transactions on Multimedia},
  year={2019},
  publisher={IEEE}
}