Division of Informatics
Forrest Hill & 80 South Bridge
MSc Thesis #9508 | |
|---|---|
| Title: | Recognizing Walking People |
| Authors: | Birbilis,A |
| Date: | 1995 |
| Presented: | |
| Keywords: | |
| Abstract: | It has been proved by other researchers that humans have the ability to recognise a familiar person, based on minimum information derived from their gait. The principal aim of this project was to investigate ways of duplicating this ability in an automated system. In order to achieve this and MLD approach was considered. Under the assumptions posed by this approach we digitalised people walking towards a steady camera. The algorithm that was developed was extracting the spatiotemporal trajectories by means of preselected points on the persons. All processing was done on a sequence of grey scale images. Reductions on the number of points under consideration were made possible with their cross-correlations. Based on the Fourier Analysis of the trajectories we derived the amplitudes of the fundamental frequency and the harmonics for each of the -reduced in number- selected points. Moreover, the amplitudes obtained for each person were tested with a Hierarchical Clustering Algorithm which illustrated that distinctions between persons are possible. Finally, the recognition task was simulated on a feed-forward neural network. |
| Download: | NO ONLINE COPY |
![[Search These Pages]](http://www.dai.ed.ac.uk/icons/dai_search.gif){kind=icon}
![[DAI Home Page]](http://www.dai.ed.ac.uk/icons/dai_home.gif){kind=icon}
![[Comment]](http://www.dai.ed.ac.uk/icons/dai_comment.gif){kind=icon}