/VSOD

The repository for IEEE TCSVT 2023 (A Video Shot Occlusion Detection Algorithm Based on the Abnormal Fluctuation of Depth Information)

Primary LanguagePythonMIT LicenseMIT

A Video Shot Occlusion Detection Algorithm Based on the Abnormal Fluctuation of Depth Information

This repository contains the code and datasets for our paper (TCSVT 2023):

A Video Shot Occlusion Detection Algorithm Based on the Abnormal Fluctuation of Depth Information
Junhua Liao, Haihan Duan, Wanbing Zhao, Kanghui Feng, Yanbing Yang, Liangyin Chen

Related Works

  • Occlusion Detection for Automatic Video Editing (ACM MM 2020) [paper] [project page]
  • A Light Weight Model for Video Shot Occlusion Detection (IEEE ICASSP 2022) [paper] [code]
  • FLAD: A Human-centered Video Content Flaw Detection System for Meeting Recordings (ACM NOSSDAV 2022) [paper]

Dataset

Method

Data preparation

Accessing this repository, utilize DPT-Large to obtain the depth map of the video.

Testing

Set the depth_path in the script and execute the following code.

python test.py

Citation

Please cite our papers if you use this code or datasets.

@inproceedings{liao2020occlusion,
  title={Occlusion Detection for Automatic Video Editing},
  author={Liao, Junhua and Duan, Haihan and Li, Xin and Xu, Haoran and Yang, Yanbing and Cai, Wei and Chen, Yanru and Chen, Liangyin},
  booktitle={Proceedings of the 28th ACM International Conference on Multimedia},
  pages={2255--2263},
  year={2020}
}

@ARTICLE{10182309,
  author={Liao, Junhua and Duan, Haihan and Zhao, Wanbing and Feng, Kanghui and Yang, Yanbing and Chen, Liangyin},
  journal={IEEE Transactions on Circuits and Systems for Video Technology}, 
  title={A Video Shot Occlusion Detection Algorithm Based on the Abnormal Fluctuation of Depth Information}, 
  year={2023},
  volume={},
  number={},
  pages={1-1},
  doi={10.1109/TCSVT.2023.3295243}}

Acknowledgments

Thanks for the support of the open-source repository (DPT, NeWCRFs, GuidedDecoding, Jun et al.) for this research.