METODE EKSTRAKSI CIRI 2DPCA PADA PENGENALAN CITRA WAJAH DENGAN MATLAB

Publication Type:

Journal Article

Authors:

Pratiwi

Source:

Jurnal Teknologi, Volume 7, Number 1 (2014)

Keywords:

2DPCA, Euclidean Distance, Face Recognition, Fiture Extraction

Abstract:

Research on face recognition have been carried out and the various methods. However, this study is still interesting to do with the development of technologies and methods used in image processing . Feature extraction method at introduction face recognition much is to identify specific facial image using measurable characteristics . In this study, researchers used a feature extraction 2 Dimensional Principal Component Analysis ( 2DPCA and using Matlab application. In this process faces read as a matrix and made the same size 112x96 pixel, and then do the conversion to grayscale.. After the feature extraction of face images 2DPCA testing was conducted using Euclidean distance. Method is the method by comparing the Euclidean distance of the test image by the training image data base that has minimal distance. Clustering of training data and test data using a 5 fold crossvalidation . Accuracy of the results obtained during experiments conducted at 10 percent face with eigenvalues 80% , 85% , 90% and 95%, the result is an average of over 96.5%.

AttachmentSize
1-5-pratiwi.pdf169.89 KB

(c) 2009 Jurnal teknologi FTI IST AKPRIND
Developed by UPT PUSKOM IST AKPRIND

Who's online

There are currently 0 users and 14 guests online.