vireoJD-MM at Activity Detection in Extended Videos

Research output: Working Papers › Preprint

View graph of relations

Author(s)

  • Fuchen Long
  • Qi Cai
  • Zhaofan Qiu
  • Yingwei Pan
  • Ting Yao

Related Research Unit(s)

Detail(s)

Original languageEnglish
PublisherarXiv
Number of pages4
Publication statusOnline published - 20 Jun 2019

Link(s)

DOI

DOI
Check@CityULib
Permanent Link https://scholars.cityu.edu.hk/en/publications/publication(6dd21082-8044-4e0b-8a3a-da4f14c0d531).html

Abstract

This notebook paper presents an overview and comparative analysis of our system designed for activity detection in extended videos (ActEV-PC) in ActivityNet Challenge 2019. Specifically, we exploit person/vehicle detections in spatial level and action localization in temporal level for action detection in surveillance videos. The mechanism of different tubelet generation and model decomposition methods are studied as well. The detection results are finally predicted by late fusing the results from each component.

Research Area(s)

  • cs.CV

Fingerprint

Publication fingerprints text

100
detection INIS
100
Activity Detection Keyphrases
100
Detection Computer Science
50
Decomposition Metho... Keyphrases
50
Comparative Analysi... Keyphrases
50
Detection Result Keyphrases
50
Vehicle Detection Keyphrases
50
Surveillance Video Keyphrases
50
Overview Analysis Keyphrases
50
Spatial Actions Keyphrases
Full Fingerprint ›

Citation Format(s)

[ RIS ] [ BibTeX ]
vireoJD-MM at Activity Detection in Extended Videos. / Long, Fuchen; Cai, Qi; Qiu, Zhaofan et al.
arXiv, 2019.

Research output: Working Papers › Preprint