Perancangan Prototype Brankas Menggunakan Sistem Pengenalan Wajah Dengan Metode Convolutional Neural Network (CNN)

Abstract

A safe-deposit box is a box used for keeping precious items. Safe-deposit boxes are designed to be difficult for people to open by force. There are various security systems that may be used in it, such as mechanical key, combinational lock, PIN, etc. However, a safe-deposit box is still prone to unpermitted access because anyone who knows the PIN or possesses the key is still able to open it. This research aims to create a safety-box prototype which has a face recognition system implemented on it to ensure no unauthorized person may access this box. Experiment is performed on three different classes, which are “erwin”, “unknown”, and “willy”. Class of “erwin” and “willy” are defined as safe owners, while “unknown” is defined as anyone who is not both owners. Classification on safe owners is considered success if the percentage output in corresponding classes is at least 90 %. Classification on “unknown” class is considered success if the result is at least 90 % or percentage on each class is lower than 90 %. Accuracy for each class is 0 %, 71.43 %, dan 100 %.