Ballet pose recognition: A bag-of-words support vector machine model for the dance training environment

Margaux Fourie, Dustin van der Haar

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

3 Citations (Scopus)

Abstract

Serious dance students are always looking for ways in which they can improve their technique by practising alone at home or a studio by using a mirror for feedback. The problem these students face is that for many ballet postures it is difficult to analyze one’s own faults. By not having guidance regarding proper positional alignment, dancers risk developing injuries and bad habits. The proposed solution is a system which recognizes the ballet position being performed by a dancer. After recognition, this research aims to work towards providing the necessary correction as feedback. The results for recognition in the system, using a Bag-of-Words approach to a Support Vector Machine classifier, showed an accuracy of 59.6%. Multiple implementations are produced and assessed in this paper. It is clearly found that the approach is feasible, however, work for improving the accuracy is required. Recommendations to improve effective pose recognition for future work are therefore discussed.

Original languageEnglish
Title of host publicationInformation Science and Applications 2018 - ICISA 2018
EditorsKuinam J. Kim, Kuinam J. Kim, Nakhoon Baek
PublisherSpringer Verlag
Pages317-325
Number of pages9
ISBN (Print)9789811310553
DOIs
Publication statusPublished - 2019
EventInternational Conference on Information Science and Applications, ICISA 2018 - Kowloon, Hong Kong
Duration: 25 Jun 201827 Jun 2018

Publication series

NameLecture Notes in Electrical Engineering
Volume514
ISSN (Print)1876-1100
ISSN (Electronic)1876-1119

Conference

ConferenceInternational Conference on Information Science and Applications, ICISA 2018
Country/TerritoryHong Kong
CityKowloon
Period25/06/1827/06/18

Keywords

  • Ballet training
  • Computer vision
  • Posture recognition
  • SVM

ASJC Scopus subject areas

  • Industrial and Manufacturing Engineering

Fingerprint

Dive into the research topics of 'Ballet pose recognition: A bag-of-words support vector machine model for the dance training environment'. Together they form a unique fingerprint.

Cite this