Identifikasi Sidik Jari pada Sistem Peminjaman Mandiri Menggunakan Support Vector Machine

Abstract

The self-service borrowing system is an innovative service that allows users to borrow laboratory equipment without direct involvement from staff. The traditional system with manual recording is considered inefficient because it is prone to recording errors, the risk of losing forms, and long queues. This study aims to develop a self-service borrowing system that integrates fingerprint biometric technology for security authentication and QR codes for automated item recording. The user identification process is carried out through feature extraction using Histogram of Oriented Gradients (HOG), which is then classified with a Support Vector Machine (SVM) algorithm based on Radial Basis Function (RBF) kernel on a Raspberry Pi device. The test results show that the SVM model successfully achieved a training accuracy of 94.7%. In direct system testing (1:N identification), the system achieved an accuracy of 94.06% with a False Rejection Rate (FRR) of 6% and a False Acceptance Rate (FAR) of 7%. All QR Code scans of items were successfully performed and directed to the database system in realtime. This system has proven effective in enhancing the security and efficiency of laboratory equipment management through service automation.