Please use this identifier to cite or link to this item:
https://ptsldigital.ukm.my/jspui/handle/123456789/475667| Title: | Multi-view gait recognition based on enhanced gait energy image and Radon transform techniques |
| Authors: | Iman Mohammed Burhan (P68482) |
| Supervisor: | Md. Jan Nordin, Assoc. Prof. Dr. |
| Keywords: | Gait energy image Radon transform Multi-view Human recognition Dissertations Academic - Malaysia |
| Issue Date: | 30-Apr-2016 |
| Description: | Biometrics is one way for the identification of individuals through their physical and behavioral characteristics. The characteristics such as iris, face, fingerprint, gait, and spoken speech. We known as biometric identifiers which are distinctive as they are attached to a person permanently. One of the biometric identifiers that has not only gained public recognition, but also well accepted as a security assessment is gait. It is a biometric modality for the identification of a human at a distance. However, gait recognition (GR) systems encounter several challenges, including viewing angles and translation variations. Hence, GR systems require the development of a robust gait representation model which is invariant in varying conditions. As such, this research study presents a gait representation model for multi-view gait recognition systems (MvGRS) based on gait energy image (GEI) and Radon Transform (RT) on human silhouettes to overcome the challenges in human recognition. In this regard, GEI is utilized for the description of gait features of binary silhouette images which are robust in multi-viewing and varied in appearances. Therefore, to ensure efficiency in gait representation, image enhancement is used to remove noises from silhouette images before the extraction of gait feature. Furthermore, the adoption of Radon Transform (RT) allows the accommodation of gait representation model with RT features and silhouette alignment. This is to overcome the difficulties in geometrical transformation such as translation, scale and rotation. Consequently robust Principal Component Analysis (PCA) and Partial Least Square (PLS) approaches are applied in the reduction of these dimension feature vectors and feature selection. Finally, the recognition of gaits is based on similarity in measurements using Euclidean distance. In order to speed up gait feature representation, the proposed method of this research study implements the multi-threading platform which supports parallel processing, hence improves the speed in gait recognition. The experiments were conducted on the public data set of CASIA B. The findings from these experiments show that the results are better in comparison with the other methods, such as Zheng's method and Kusakunniran's method. Thus, this indicates that the proposed method for gait recognition can outperform the existing methods in gait recognition.,Certification of Master's/Doctoral Thesis" is not available |
| Pages: | 79 |
| Publisher: | UKM, Bangi |
| Appears in Collections: | Faculty of Information Science and Technology / Fakulti Teknologi dan Sains Maklumat |
Files in This Item:
There are no files associated with this item.
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.