Implementasi Speech Recognition dan Face Detection pada Sistem Pelayanan Informasi Aplikasi Satu Atap (ASA) Universitas Sumatera Utara

Abstract

In this increasingly advanced era, it has become possible to simplify human tasks, including in information services. At the University of Sumatera Utara, there is an information service unit known as the One-Stop Application (ASA), which serves as a hub for correspondence interactions between students, study programs, and faculties. The ASA unit has an online correspondence system application; however, many students still visit the ASA unit directly to inquire about their correspondence needs. Therefore, an information service system was developed to assist students in obtaining information about the documents they need without requiring assistance from staff. To improve efficiency and convenience, an information service system based on speech recognition and face detection was developed. Speech recognition functions as a voice interaction interface, allowing students to inquire about document requirements without direct interaction with staff. This technology is equipped with face detection to ensure that the system does not operate continuously—if no person is detected in front of the camera, the system remains inactive. In this research, face detection was tested with 20 individuals, yielding an average detection time of 1.45 seconds, demonstrating the system's ability to identify faces quickly and accurately. Speech recognition was tested using complete sentences or key phrases, and the results showed that the system could recognize sentences with high accuracy in the given scenarios. This research demonstrates that the integration of speech recognition and face detection in the ASA unit can help students access document-related information quickly, securely, and efficiently. This system is expected to be implemented more broadly in other academic administrative services to enhance service quality within the university environment.