English

Trimmed Action Recognition, Dense-Captioning Events in Videos, and Spatio-temporal Action Localization with Focus on ActivityNet Challenge 2019

Computer Vision and Pattern Recognition 2019-06-18 v1

Abstract

This notebook paper presents an overview and comparative analysis of our systems designed for the following three tasks in ActivityNet Challenge 2019: trimmed action recognition, dense-captioning events in videos, and spatio-temporal action localization.

Keywords

Cite

@article{arxiv.1906.07016,
  title  = {Trimmed Action Recognition, Dense-Captioning Events in Videos, and Spatio-temporal Action Localization with Focus on ActivityNet Challenge 2019},
  author = {Zhaofan Qiu and Dong Li and Yehao Li and Qi Cai and Yingwei Pan and Ting Yao},
  journal= {arXiv preprint arXiv:1906.07016},
  year   = {2019}
}

Comments

arXiv admin note: substantial text overlap with arXiv:1807.00686, arXiv:1710.08011