chechao/hollywood-2-actions

This is the Hollywood-2 datasets from DHF1K for Video Saliency Detection, all credits belong to them. Since their Google Drive link is dead. I downloaded from Baidu Disk and reuploaded it here for anyone who doesn't have access to Baidu Disk.

Please cite their work if you found the dataset useful

@InProceedings{Wang_2018_CVPR,
author = {Wang, Wenguan and Shen, Jianbing and Guo, Fang and Cheng, Ming-Ming and Borji, Ali},
title = {Revisiting Video Saliency: A Large-Scale Benchmark and a New Model},
booktitle = {The IEEE Conference on Computer Vision and Pattern Recognition},
year = {2018}
}

@ARTICLE{Wang_2019_revisitingVS, 
author={W. {Wang} and J. {Shen} and J. {Xie} and M. {Cheng} and H. {Ling} and A. {Borji}}, 
journal={IEEE Transactions on Pattern Analysis and Machine Intelligence}, 
title={Revisiting Video Saliency Prediction in the Deep Learning Era}, 
year={2019}, 
}