Author: Liu, Peng
Title: Action recognition using deep learning models
Degree: M.Sc.
Year: 2019
Subject: Hong Kong Polytechnic University -- Dissertations
Computer vision
Pattern recognition systems
Image processing -- Digital techniques
Department: Faculty of Engineering
Pages: vii, 42 pages : color illustrations
Language: English
Abstract: Action recognition has become a research hotspot in the field of computer vision in view of its wide application in the fields of video tracking, motion analysis, auxiliary medicine, virtual reality and human-computer intelligent interaction. Action recognition refers to the act of classifying an action that is present in a given video and action detection involves locating actions of interest in space or time. Although many methods of action recognition have been proposed so far, action recognition is still full of challenges. A major difficulty is that the model can not only detect targets and actions from the background, but also accurately identify changes in action diversity. In response to this problem, the traditional model will be hindered. This dissertation proposes using skeleton data for action recognition. In the experiment, the resolution of the video was reduced. With the help of openpose, two usages are introduced: combining 3D CNN and skeleton data and adding skeleton information into initial videos. I have conducted two sets of experiments separately: traditional 3D CNN, combination of 3D CNN and skeleton and combined videos. Compared with 3D CNN, the other two methods have a small promotion. That shows the skeleton data works a little. For some different classes of video performance, we analyzed this result to some extent. Overall, when skeleton data is used as auxiliary information, the performance will be improved to a small extent. But this kind of improvement requires certain conditions.
Rights: All rights reserved
Access: restricted access

Files in This Item:
File Description SizeFormat 
991022270856903411.pdfFor All Users (off-campus access for PolyU Staff & Students only)1.34 MBAdobe PDFView/Open

Copyright Undertaking

As a bona fide Library user, I declare that:

  1. I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
  2. I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
  3. I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.

By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.

Show full item record

Please use this identifier to cite or link to this item: