Full metadata record
| DC Field | Value | Language |
|---|---|---|
| dc.contributor | Faculty of Engineering | en_US |
| dc.contributor.advisor | Lam, Kin Man (EIE) | - |
| dc.creator | Gong, Chenhui | - |
| dc.identifier.uri | https://theses.lib.polyu.edu.hk/handle/200/10094 | - |
| dc.language | English | en_US |
| dc.publisher | Hong Kong Polytechnic University | - |
| dc.rights | All rights reserved | en_US |
| dc.title | Action recognition : improvements for spatio-temporal laplacian pyramid coding | en_US |
| dcterms.abstract | In computer vision research, human action analysis in videos has attracted more and more attention. Many recognition algorithms have been reported recently. This dissertation focuses on a study of the holistic representation: spatio-temporal Laplacian pyramid coding (STLPC), and improves its shortcomings. Compared with sparse representation which detects local interest points, this method can encode more visual information. STLPC considers a video sequence as a 3-D volume, which is called a video volume, and the spatio-temporal features are extracted from it. Through the establishment of pyramid model and the use of 3-D Gabor filters, the information which comes from different scales and orientations is obtained. The max pooling method is then applied to preserve useful information. STLPC can generate a representation with action information and the structure is also included. However, holistic representations also have some shortcomings. They are sensitive to occlusions and background variations. These problems can be overcome and the final result can be improved through a background subtraction method, which is called ViBe+, to segment out human body from a video sequence. ViBe+ is a pixel-level model. It can be used to extract background and detect foreground. This segmentation method is better than several well-known algorithms. The proposed method is evaluated on the KTH and UCF datasets. Experiment results show that the recognition rates of the proposed method are better than many other methods. | en_US |
| dcterms.extent | vi, 66 pages : color illustrations | en_US |
| dcterms.isPartOf | PolyU Electronic Theses | en_US |
| dcterms.issued | 2019 | en_US |
| dcterms.educationalLevel | M.Sc. | en_US |
| dcterms.educationalLevel | All Master | en_US |
| dcterms.LCSH | Hong Kong Polytechnic University -- Dissertations | en_US |
| dcterms.LCSH | Computer vision | en_US |
| dcterms.LCSH | Pattern recognition systems | en_US |
| dcterms.LCSH | Image processing -- Digital techniques | en_US |
| dcterms.accessRights | restricted access | en_US |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| 991022270858203411.pdf | For All Users (off-campus access for PolyU Staff & Students only) | 3.9 MB | Adobe PDF | View/Open |
Copyright Undertaking
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- 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.
- 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.
Please use this identifier to cite or link to this item:
https://theses.lib.polyu.edu.hk/handle/200/10094