Action recognition by pairwise proximity function support vector machines with dynamic time warping kernels

Mohammad Ali Bagheri*, Qigang Gao, Sergio Escalera

*Kontaktforfatter

Publikation: Bidrag til bog/antologi/rapport/konference proceedingKonferenceartikel i proceedingForskningpeer review

Abstract

In the context of human action recognition using skeleton data, the 3D trajectories of joint points may be considered as multidimensional time series. The traditional recognition technique in the literature is based on time series dis(similarity) measures (such as Dynamic Time Warping). For these general dis(similarity) measures, k-nearest neighbor algorithms are a natural choice. However, k-NN classifiers are known to be sensitive to noise and outliers. In this paper, a new class of Support Vector Machine that is applicable to trajectory classification, such as action recognition, is developed by incorporating an efficient time-series distances measure into the kernel function. More specifically, the derivative of Dynamic Time Warping (DTW) distance measure is employed as the SVM kernel. In addition, the pairwise proximity learning strategy is utilized in order to make use of non-positive semi-definite (PSD) kernels in the SVM formulation. The recognition results of the proposed technique on two action recognition datasets demonstrates the our performance of our methodology compared to the state-of-the-art methods. Remarkably, we obtained 89% accuracy on the well-known MSRAction3D dataset using only 3D trajectories of body joints obtained by Kinect.

OriginalsprogEngelsk
TitelAdvances in Artificial Intelligence - 29th Canadian Conference on Artificial Intelligence, Canadian AI 2016, Proceedings
RedaktørerRichard Khoury, Christopher Drummond
Antal sider12
ForlagPhysica-Verlag
Publikationsdato2016
Sider3-14
ISBN (Trykt)9783319341101
DOI
StatusUdgivet - 2016
Udgivet eksterntJa
Begivenhed29th Canadian Conference on Artificial Intelligence, AI 2016 - Victoria, Canada
Varighed: 31 maj 20163 jun. 2016

Konference

Konference29th Canadian Conference on Artificial Intelligence, AI 2016
Land/OmrådeCanada
ByVictoria
Periode31/05/201603/06/2016
Sponsor(CAIAC), Canadian Artificial Intelligence Association, Cory Butz (University of Regina)
NavnLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Vol/bind9673
ISSN0302-9743

Bibliografisk note

Publisher Copyright:
© Springer International Publishing Switzerland 2016.

Fingeraftryk

Dyk ned i forskningsemnerne om 'Action recognition by pairwise proximity function support vector machines with dynamic time warping kernels'. Sammen danner de et unikt fingeraftryk.

Citationsformater