AbstrakTerdapat banyak metode yang digunakan untuk mengenali identitas seseorang, misalkan nomor unik, kartu identitas dan sandi rahasia. Kekurangan metode-metode tersebut antara lain, kartu dapat hilang, nomor unik dan sandi rahasia dapat terlupakan. Salah satu solusi untuk masalah ini adalah sistem identifikasi seseorang berdasarkan metode biometrik jenis fisiologis. Penelitian ini merancang sebuah sistem untuk mengidentifikasi wajah. Citra wajah diambil menggunakan kamera web kemudian diekstraksi cirinya dengan metode local binary pattern (LBP). Ciri wajah yang diperoleh diklasifikasi menggunakan support vector machine (SVM). Model terbaik SVM dibangun berdasarkan validasi silang grid search. Kernel linier terbaik dibentuk dengan LBP 8,2 u2 dan parameter C = 10 5 . Kernel radial basis function (RBF) terbaik dicapai dengan LBP 16,2 u2 dan parameter C = 10 3 dan γ = 10 2 . Berdasarkan pengujian terhadap keseluruhan citra wajah, akurasi kedua kernel adalah 96,0%. Pada pengujian lima ekspresi wajah dengan SVM kernel linier, akurasi 100,0% diperoleh untuk ekspresi sedih, netral dan mata tertutup. Sedangkan SVM kernel RBF menghasilkan akurasi 100,0% untuk ekspresi terkejut, netral dan mata tertutup. Hasil pengujian tersebut menunjukkan sistem pengenalan wajah yang dirancang telah berfungsi baik. Kata kunci: pengenalan wajah, local binary pattern, support vector machine, validasi silang grid search AbstractThere are many methods to recognize a person's identity, e.g. unique number, ID card, and password. The methods have some weaknesses, e.g. ID can be lost, unique number and password can be forgotten. One of solution to these problems is recognition system based on physiological biometric. This Research designs a system to identify human face. Face images are taken using webcam and then local binary pattern (LBP) is employed for feature extraction. The obtained face features are classified using support vector machine (SVM). The best models are built based on grid search cross validation. The best linier kernel is constructed with LBP 8,2 u2 and parameter C = 10 5 . The best radial basis function (RBF) kernel is achieved with LBP 16,2 u2 and the parameters C = 10 3 and γ = 10 2 . Test to all images using both kernels produces an accuracy of 96.0%. In test based on five facial expressions shows that linier kernel achieves an accuracy of 100,0% in sad, neutral and closed eyes and RBF kernel obtains an accuracy of 100,0% for shocked, neutral and closed eyes. The results of these tests show that the face recognition system works well. keywords: face recognition, local binary pattern, support vector machine, grid search cross validation.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.