Pengenalan Jenis Kelamin Berbasis Kernel Principal Component Analysis

Achmad Rizal

Abstract

Gender Recognition  adalah salah satu penelitian di bidang biometrik dan computer vision yang cukup popular. Gender Recognition adalah pengembangan dari Face Recognition, Gender Recognition dapat mengklasifikasikan citra menjadi 2 kelas yaitu perempuan dan laki-laki. Penelitian ini menggunakan 400 citra, 200 citra perempuan dan 200 citra laki-laki dan memakai database JAVE yang telah teruji sebelumnya. Ada 2 tahapan penting dalam penelitian ini, tahap pertama ektraksi fitur menggunakan metode Kernel Principal Component Analysis. Metode kernel dapat membuat representasi data pada ruang kernel dengan menggunakan fungsi non-linear dan kernel yang digunakan adalah linier kernel. Tahap kedua adalah pengukuran jarak kemiripan citra testing terhadap citra training menggunakan metode 2D Correlation Coefficient yang bekerja dengan cara mengalikan data training dan data testing kemudian membagi dengan hasil akar kuadratnya. Metode kernel berjalan cukup baik karena memperoleh akurasi lebih tinggi dari metode dari penelitian sebelumnya yang telah digunakan. Akurasi tertinggi yang dihasilkan pada penelitian ini mencapai 92%.

Kata Kunci : Face Recognition, Gender Recognition, Kernel Principal Component Analysis, Correlation Coefficient.

ABSTRACT

Gender Recognition is one of the research areas in the field of biometrics and computer vision that are quite popular. Gender Recognition is the development of Face Recognition. Gender Recognition can classify the image into two classes, namely women and men. This thesis used 400 images, 200 images and 200 images of women men and used the JAVE database that had been tested previously. There are two important stages in this research. The first phase is a feature extraction by using Kernel Principal Component Analysis method. Kernel methods can make a representation of data in kernel space using a non-linear function and the kernel used is a linear kernel. The second stage is the measurement of the distance to the testing image similarity of training images by using 2D Correlation Coefficient method that works by multiplying the training data and testing data and then dividing them by the results of the square root. Kernel methods work quite well in which the accuracy is higher than previous methods of the research that have been used. Produced the highest accuracy in this study reached 92%.

Keywords: Face Recognition, Gender Recognition, Kernel Principal Component Analysis, Correlation Coefficient.

Keywords

Face Recognition, Gender Recognition, Kernel Principal Component Analysis, Correlation Coefficient

Full Text:

PDF

References

Purnomo, M.H., dan Muntasa, A., Konsep Pengolahan Citra Digital dan Ekstraksi Fitur. Yogyakarta: Graha Ilmu. 2010.

Riko, A.S., Pengenalan Wajah Menggunakan Metode Fisherface. Tugas Akhir. Universitas Kristen Maranatha, Bandung. 2007.

Quan, W., Kernel Principal Component Analysis and its Aplication in Face Recognition and Active Shape Models. Thesis. Renselaer Polytechnic. USA. 2012.

Kekre, H.B., Thepade, S.D., Chopra, T., Face and Gender Recognition Using Principal Component Analysis.(IJCSE) International Journal on Computer Science and Engineering, 2: 1-6. 2010.

Eugene, K., Yen., dan Roger G.J., The In Effectiveness of The Correlation Coefficient for Image Comparisons. Thesis. Los Alamos National Laboratory. Los Alamos, New Mexico.

Ferdiansah, A. Gender Recognition Menggunakan Metode Principal Component Analysis Dikombinasikan Dengan Local Binary Pattern. Bangkalan:Universitas Trunojoyo Madura. 2013.

DOI

https://doi.org/10.21107/rekayasa.v8i1.5356

Metrics

Refbacks

  • There are currently no refbacks.


Copyright (c) 2016 Achmad Rizal

Creative Commons License
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.