Deep learning for action recognition and prediction

Qiuhong Ke

Research output: ThesisDoctoral Thesis

833 Downloads (Pure)

Abstract

This thesis proposes novel deep leaning solutions for real-world applications, including action recognition and prediction from videos. The first part leverages Convolutional Neural Networks (CNN) to learn deep spatial-temporal information of skeleton sequences for action recognition. The second part presents three new methods based on CNN and Recurrent Neural Networks for action prediction. The last part of this thesis presents a new feature extraction method based on CNN for better person re­identification. Extensive experiments have shown the superiority of the proposed methods compared to state of-the-art methods for action recognition, action prediction and person re-identification.
Original languageEnglish
QualificationDoctor of Philosophy
Awarding Institution
  • The University of Western Australia
Supervisors/Advisors
  • Bennamoun, Mohammed, Supervisor
  • Boussaid, Farid, Supervisor
  • Sohel, Ferdous, Supervisor
  • An, Senjian, Supervisor
Thesis sponsors
Award date25 Jul 2018
DOIs
Publication statusUnpublished - 2018

Fingerprint

Dive into the research topics of 'Deep learning for action recognition and prediction'. Together they form a unique fingerprint.

Cite this