<![CDATA[ABSTRAK Perkembangan teknologi di era digitalisasi telah memberikan dampak signifikan pada berbagai sektor, termasuk pendidikan tinggi. Inovasi teknologi gencar dilakukan untuk mendukung kegiatan perkuliahan, salah satunya adalah sistem presensi mahasiswa. Beberapa diantaranya yang sudah diterapkan adalah teknologi RFID pada Kartu Tanda Mahasiswa dan scan kode QR. Namun keduanya masih memiliki celah kekurangan seperti risiko terjadinya kehilangan KTM, penyalahgunaan kode QR atau bahkan fenomena titip absen. Penggunaan sistem biometri seperti face recognition menjadi alternatif inovasi untuk meningkatkan keabsahan presensi mahasiswa. Penelitian ini mengimplementasikan teknologi face recognition secara real time pada sistem presensi mahasiswa berbasis web dengan mengombinasikan metode Single Shot Multibox Detector (SSD) dan Local Binary Pattern Histogram (LBPH). Metode SSD digunakan sebagai pendeteksi wajah (face detector) dan LBPH sebagai pengenal wajahnya (face recognizer). Penelitian ini melibatkan pengujian akurasi dalam mendeteksi dan mengenali wajah mahasiswa berdasarkan parameter jarak, jumlah wajah dalam satu frame dan posisi wajah. Pada pengujian deteksi wajah untuk parameter jarak radius 30 cm hingga 100 cm dan posisi wajah diperoleh akurasi 100%. Pengujian pengenalan wajah berdasarkan posisi wajah memperoleh akurasi mencapai 85% dan presisi 87%, pengenalan wajah berdasarkan jarak memperoleh akurasi sebesar 85% dan presisi sebesar 86%. ABSTRACT The development of technology in the era of digitalization has had a significant impact on various sectors, including higher education. Technological innovations have been actively pursued to support academic activities, one of which is the student attendance system. Some of the technologies that have been implemented include RFID technology on Student Identification Cards and QR code scanning. However, both methods still have shortcomings, such as the risk of losing the Student Identification Card, QR code misuse, or even the phenomenon of proxy attendance. The use of biometric systems, such as face recognition, has become an alternative innovation to enhance the validity of student attendance. This research implements real-time face recognition technology in a web-based student attendance system by combining the Single Shot Multibox Detector (SSD) and Local Binary Pattern Histogram (LBPH) methods. The SSD method is used as the face detector, and LBPH is used as the face recognizer. This research involves testing accuracy in detecting and recognizing student faces based on distance parameters, number of faces in one frame and face position. In face detection testing for radius distance parameters of 30 cm to 100 cm and face position, 100% accuracy was obtained. Testing facial recognition based on facial position obtained accuracy of 85% and precision of 87%, facial recognition based on distance obtained accuracy of 85% and precision of 86%.]]>
Copyrights © 2024
