JavaScript is disabled for your browser. Some features of this site may not work without it.
| dc.contributor.author | ADMASU, YEHULE | |
| dc.date.accessioned | 2022-03-24T06:39:13Z | |
| dc.date.available | 2022-03-24T06:39:13Z | |
| dc.date.issued | 2021-09 | |
| dc.identifier.uri | http://ir.bdu.edu.et/handle/123456789/13244 | |
| dc.description.abstract | Due to fast growing technologies, today’s world becomes more competitive than ever before. Face recognition is the most interesting and grateful application of pattern recog-nition, image analysis and going on significant attention in security system. It has been used as a dominant method for security purpose which can be considered as key public safety issue for modern society. In this work first we prepare data set by split into train and test set up until by performing recognition and classification due to the selection of the input data, then finally we computing the accuracy of system. And also, in this work we can perform face recognition and detection in two-dimensional aspects using different algorithms. However, without test image applying we can not perform face recognition and detection. To achieve this, several algorithms have been implemented, such as Linear Discriminant Analysis (LDA), K-Nearest Neighbors (KNN), and Principal Component Analysis (PCA) and Voila-Jone(VJ). Even though, these applications have been success-fully applied and showed good results which may fail to accurately represent the face due many factors. But, some of this applications have their own advantages. PCA is an easier approach used to distinguish the face in the given image and it is considered as a suitable approach. On the other hand, Voila-Jone algorithm is a better algorithm that is used in face detec-tion mechanism, and it is capable of detecting a face in two dimensional aspects. In this thesis we have made a comparative study of output of different algorithms. The study was conducted and tested based some of the standard face data sets, such as ORL, Yale and Face94 data sets. Also, we apply different number of testing and training images are used for performance evaluation. When 45 images were tested, PCA has shown an accuracy of 97.78%, while Voila-Jone algorithm has shown 100% of detection accuracy. In addition, 40 images were tested, LDA and KNN have shown an accuracy of 100% and 98.5% respectively, but Voila-Jone algorithm achieved 97.5% of detection accuracy. LDA is better results in the other methods in terms of accuracy. Key Words: PCA, face recognition, Voila-Jone Algorithm, LDA, KNN | en_US |
| dc.language.iso | en_US | en_US |
| dc.subject | Communication System Engineering | en_US |
| dc.title | Performance Analysis of Two-Dimensional Face Detection and Recognition Algorithms | en_US |
| dc.type | Thesis | en_US |
