Updates

MovieNet Official Website is online now!

Introduction

movienet-tools is an open source movie analysis toolbox based on PyTorch. It's part of MovieNet project maintained by MovieNet Team from MMLab, CUHK. And it is also one of OpenMMLab projects.

Features

• Basic video processing tools.
• Holistic semantic video feature extractors.
• All-in-one movie info web crawler.

Installation

Please refer to INSTALL.md for installation and dataset preparation. Pretrained models and dataset are also explanined here.

Get Started

Please see GETTING_STARTED.md for the wiki of movienet data and the basic usage of this toolbox.

Acknowledgement

The structure of movienet-tools follows that of codebased in openmmlab. The part of character detection are modified from mmdetection and the part of action feature extraction are modified from mmaction. Many thanks to these open-source codebases.

Citation

If you use MovieNet dataset or this toolbox or benchmarks in your research, please cite this project.

@inproceedings{huang2020movie,
	title={MovieNet: A Holistic Dataset for Movie Understanding},
	author={Huang, Qingqiu and Xiong, Yu and Rao, Anyi and Wang, Jiaze and Lin, Dahua},
	booktitle = {Proceedings of the European Conference on Computer Vision (ECCV)}, 
	year={2020}
}