Skip to main navigation Skip to search Skip to main content

Action detection with improved dense trajectories and sliding window

  • Stony Brook University

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

16 Scopus citations

Abstract

In this paper we describe an action/interaction detection system based on improved dense trajectories [19], multiple visual descriptors and bag-of-features representation. Given that the actions/interactions are not mutual exclusive, we train a binary classifier for every predefined action/interaction. We rely on a non-overlapped temporal sliding window to enable the temporal localization. We have tested our system in ChaLearn Looking at People Challenge 2014 Track 2 dataset [1,2]. We obtained 0.4226 average overlap, which is the 3rd place in the track of the challenge. Finally, we provide an extensive analysis of the performance of this system on different actions and provide possible ways to improve a general action detection system.

Original languageEnglish
Title of host publicationComputer Vision - ECCV 2014 Workshops, Proceedings
EditorsMichael M. Bronstein, Carsten Rother, Lourdes Agapito
PublisherSpringer Verlag
Pages541-551
Number of pages11
ISBN (Electronic)9783319161778
DOIs
StatePublished - 2015
Event13th European Conference on Computer Vision, ECCV 2014 - Zurich, Switzerland
Duration: Sep 6 2014Sep 12 2014

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume8925

Conference

Conference13th European Conference on Computer Vision, ECCV 2014
Country/TerritorySwitzerland
CityZurich
Period09/6/1409/12/14

Keywords

  • Action detection
  • Action recognition
  • Dense trajectories
  • Video analysis

Fingerprint

Dive into the research topics of 'Action detection with improved dense trajectories and sliding window'. Together they form a unique fingerprint.

Cite this