Please use this identifier to cite or link to this item:
http://dspace.unimap.edu.my:80/xmlui/handle/123456789/9886
Title: | Feature-based face recognition system using utilized artificial neural network |
Authors: | Chai, Tong Yuen |
Keywords: | Face recognition Algorithms Artificial intelligent (AI) Computer vision Artificial neural network Facial features cropping |
Issue Date: | 2009 |
Publisher: | Universiti Malaysia Perlis |
Abstract: | This project aims to reduce the effect of critical conditions such as excessive illumination, facial expressions, hairstyles, beard and moustache which have affected the performance of face recognition since ages ago. The main contributions of this project are the automatic algorithms for mouth detection, facial features cropping and face classification. First, the algorithm will detect a human face and irises. Second, the mouth region is estimated by using geometric calculation based on the irises positions. A proposed algorithm which combines RGB color map and corner detection techniques will detect the mouth corners. Then, the proposed features cropping system will crop the detected iris and mouth automatically. These features are fed into the backpropagation neural network. The proposed architecture contains two neural networks. The second network merges the results from template matching and first neural network to reduce wrong recognition rate and improve the performance of neural network. The proposed automatic feature-based face recognition system has efficiency more than 95% under the stated critical conditions. All the experiment results are studied to prove the quality and uniqueness of this research. |
URI: | http://dspace.unimap.edu.my/123456789/9886 |
Appears in Collections: | School of Mechatronic Engineering (Theses) |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
Page 1-24.pdf | This item is protected by original copyright | 77.77 kB | Adobe PDF | View/Open |
Full Text.pdf | Access is limited to UniMAP community | 3.75 MB | Adobe PDF | View/Open |
Items in UniMAP Library Digital Repository are protected by copyright, with all rights reserved, unless otherwise indicated.