Analisis tingkat kekasaran Citra Foto Kulit Wajah dengan DCT dan Operator Tepi

ROCHANI, Dr.Ir. Thomas Sri Widodo, DEA

2005 | Tesis | S2 Teknik Elektro (Magister Teknik Instrumentasi)

Kulit wajah seorang wanita pada umumnya dianggap sebagai cerminan kepribadian dan para wanita mendambakan kulit wajah yang halus. Para ahli kecantikan dan dokter kulit berusaha menciptakan obat-obatan yang dapat memperhalus kulit wajah wanita. Sayangnya ukuran kekasaran kulit wajah belum dibakukan, mereka biasa menggunakan citra digital foto wajah untuk meneliti kehalusan kulit. Proses citra kulit wajah dengan menggunakan alih ragam cosinus diskrit (Discrete Cosine Transform = DCT) dan operator tepi. Penelitian pertama dengan DCT pada ruang warna LUX (Logarithmic Hue Extension), YCrCb, dan HLS. Ke tiga ruang warna yang paling cocok untuk menganalisis kekasaran kulit wajah adalah ruang warna LUX yang mempunyai tiga variabel (L, U, dan X).Variabel U (hue) yang paling cocok untuk menganalisis citra kulit wajah, karena varibel hue dalam ruang warna LUX lebih mengeksploitasi spektrum merah dan relasi logaritmik pada LUX lebih representatif dari pada relasi linear ruang warna lainnya dalam menganalisis spektrum warna kulit. Pengambilan keputusan kriteria kekasaran citra kulit wajah berpedoman pada nilai varian variabel hue. Penelitian ke dua dengan deteksi tepi yang terdiri atas tujuh operator yaitu : Sobel, Prewitt, Isotropic, Kompas-1, Kompas-2 (Kirsch), Kompas-3, dan Kompas-4. Pengambilan keputusan untuk menganalisis citra kulit wajah berpedoman pada nilai rasio terkecil dari ke tujuh operator. DCT dan operator tepi dapat digunakan untuk menganalisis tingkat kekasaran citra kulit wajah.

Generally a female facial skin is considered as a personality reflection and enable have a strong willingness to have a smooth facial skin. On the other hand, beautician as well as dermatologist have tried to enable drugs which does able to keep female facial skin smoothness. Unfortunately, the facial skin smoothness has not been standardization. Dermatologist usually use digital image to measure facial skin smoothness. The facial skin is many containing red color spectral. The facial skin classified based on four criterion as very rough, rough, smooth, and very smooth. The facial skin image can be processed by the Discrete Cosine Transforms (DCT). The first observation by DCT is LUX color room (For Logarithmic Extension Hue), three the variables as L, U, and X. The variables which most suited for the facial skin image analysis is come near with the variable of hue, because the variable of hue in LUX color room more exploitation of red spectrum. The variables of U and of X the work to primary of the color that is squeezing, green, and blue. The face skin of image DCT AC analysis searching for variance value. The decision based AC coefficient hue variance value. The second observation by edge detected is searching for ratio value between amount of pixel which is edge definition and pixel all. The processing edge detected with the seven operator approach is (Sobel, Prewitt, Isotropic, Compass-1, Compass-2 (Kirsch), Compas-3, and Compass-4). The decision based is smallest ratio value from the seven operator. DCT and edge detected can be used for the facial skin image analysis

Kata Kunci : Citra Digital, Foto Wajah, DCT, Deteksi Tepi, Face Skin, DCT, LUX, Edge Detection.