Please use this identifier to cite or link to this item: http://repository.ipb.ac.id/handle/123456789/15988
Title: Face Recognition Using VFI5 Algorithm with Wavelet Transformation.
Authors: Pramitasari, Noviana
Issue Date: 2009
Publisher: IPB (Bogor Agricultural University)
Abstract: Biometrik system is the introduction of the system to identify patterns that physiological characteristics a person with determining the autenticity of a specific psysiolagical and/or behavioral characteristic posessed by that person. Not all physiological characteristics can be used on the system biometrik, several characteristics that must be fulfilled that is universal, distinctiveness, permanent, and collectability. Face is one of the physiological characteristics that can not be falsified, therefore this research using biometrik face. Wavelet is an image processing method that can extraction features and the features that are important will not be lost when the image dimension reduction. Wavelet transformation of the image will be used as an input system in the face recognition of this research and mother wavelet used was the Haar wavelet. Method face recognition used of this research is the algorithm Voting Feature Interval (VFI5). VFI5 algorithm is an algorithm that represents the description of a concept by a set of interval values of the features or attributes. Phases of the training VFI5 algorithm produce the intervals, and each features a manner resulting image data represented by pixels on each element of data. Face image used in this research is measuring 92 × 112 pixels. The image of the face will have a wavelet transformation so that will be quarter of the original dimensions, the transformation level 1 produce image measure 46 × 56, the transformation level 2 produce image measure 23 × 28, and so on until level 6 produce image measure 2 × 2. In this research using training data 6 face images and testing data 4 face images. Rank accuracy the wavelet decomposition level increased from level 2 to level 6, different with level 1 if comparing level 2 is lower accuracy. In first level the accuracy is 72,5%, 90% for second level, 85% for third level, 80% for fourth level, 68% for fifth level, and 30% for sixth level. Keywords : image prosessing, face recognition, wavelet transformation, VFI5 algorithm.
URI: http://repository.ipb.ac.id/handle/123456789/15988
Appears in Collections:UT - Computer Science

Files in This Item:
File Description SizeFormat 
abstract G09npr.pdf
  Restricted Access
Abstract294.37 kBAdobe PDFView/Open
abstract G09npr.ps
  Restricted Access
Postscript454.63 kBPostscriptView/Open
cover G09npr.pdf
  Restricted Access
Cover375.13 kBAdobe PDFView/Open
Daftar Pustaka G09npr.pdf
  Restricted Access
Daftar Pustaka302.06 kBAdobe PDFView/Open
G09npr.pdf
  Restricted Access
Full Text4.14 MBAdobe PDFView/Open
Lampiran G09npr.pdf
  Restricted Access
Lampiran3.7 MBAdobe PDFView/Open
Pendahuluan G09npr.pdf
  Restricted Access
BAB I304.51 kBAdobe PDFView/Open
Tinjuan Pustaka.pdf
  Restricted Access
BAB II340.04 kBAdobe PDFView/Open
Metodologi Penelitian G09npr.pdf
  Restricted Access
BAB III466.5 kBAdobe PDFView/Open
Pembahasan G09npr.pdf
  Restricted Access
BAB IV472.48 kBAdobe PDFView/Open
Kesimpulan G09npr.pdf
  Restricted Access
BAB V313.86 kBAdobe PDFView/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.