Awesome

Holistic Video Understanding Dataset

This repository contains the Holistic Video Understanding Dataset. For more information regarding the dataset please check our Paper and Supplementary Material.

Paper

Large Scale Holistic Video Understanding, ECCV 2020

If you use the HVU dataset, please cite our paper.

@inproceedings{hvu,
  title={Large Scale Holistic Video Understanding},
  author={Diba, Ali and Fayyaz, Mohsen and Sharma, Vivek and Paluri, Manohar and Gall, J{\"u}rgen and Stiefelhagen, Rainer and Van Gool, Luc},
  booktitle={European Conference on Computer Vision},
  pages={593--610},
  year={2020},
  organization={Springer}
}

Challenge

The submission server of the "First International Challenge on Holistic Video Understanding" is now publicly open. The winners will be announced at our HVU #CVPR2021 workshop. Submission server: https://competitions.codalab.org/competitions/29546

Usage

We store the video IDs from Youtube and their annotations in CSV format.

Dataset Downloader

Check HVU downloader instruction for easy way to download HVU.

Accessing Missing Videos and Test Videos

To access the Test videos and missing videos, please fill this form to obtain the dataset.