Analisis Kinerja Smart Door Hybrid Haar Cascade dan ArcFace pada Raspberry
Abstract
Implementing biometric security systems on deep learning devices faces a major challenge in balancing identity verification accuracy with computational resource efficiency. This study presents a performance analysis of a Raspberry Pi 5-based Smartdoor system integrating the detection speed of Haar Cascade with the recognition accuracy of ArcFace. System performance was evaluated based on RAM usage, CPU load, FPS stability, and access Success Rate parameters. Empirical evaluation results indicate that integrating Deep learning ArcFace increased RAM usage by 33.7% and CPU load from 33% to 53%. However, due to the processing capacity of the Raspberry Pi 5, the system maintained stable real-time performance with an average of 18.3 FPS. In terms of security, the Hybrid method proved superior with an access success rate of 73.7%, surpassing the conventional Haar Cascade method which only reached 68.4%. This study concludes that the Hybrid method is a viable solution for home security systems, where the increased computational load is justified by a significant improvement in identity verification reliability.
Keyword: Raspberry Pi 5; Smartdoor; Haar Cascade; ArcFace; Computational Performance.
Abstrak
Implementasi sistem keamanan biometrik pada perangkat deep learning menghadapi tantangan utama dalam menyeimbangkan akurasi verifikasi dengan efisiensi sumber daya. Penelitian ini menyajikan analisis kinerja sistem Smartdoor berbasis Raspberry Pi 5 yang mengintegrasikan kecepatan deteksi Haar Cascade dengan akurasi pengenalan wajah ArcFace. Kinerja sistem dievaluasi berdasarkan parameter penggunaan RAM, beban CPU, stabilitas FPS, dan tingkat keberhasilan akses. Hasil evaluasi empiris menunjukkan bahwa integrasi Deep learning ArcFace meningkatkan penggunaan RAM sebesar 33,7% dan beban CPU dari 33% menjadi 53%. Namun, berkat kapasitas pemrosesan Raspberry Pi 5, sistem mampu mempertahankan stabilitas kinerja real-time dengan rata-rata 18,3 FPS. Dari segi keamanan, metode Hybrid terbukti lebih unggul dengan akurasi pengenalan wajah sebesar 73,7%, melampaui metode konvensional Haar Cascade yang hanya mencapai 68,4%. Penelitian ini menyimpulkan bahwa metode Hybrid merupakan solusi yang layak untuk sistem keamanan rumah, di mana peningkatan beban komputasi terbayar dengan peningkatan reliabilitas verifikasi identitas yang signifikan.
Keywords
References
D. Y. Wijaya and A. Yulianto, “Prototype of smartdoor using RFID technology with Internet of Things (IoT),” in Proc. Conf. Management, Business, Innovation, Education and Social Science (COMBINES), vol. 1, no. 1, pp. 196–204, 2021.
L. Qinjun, C. Tianwei, Z. Yan, and W. Yuying, “Facial Recognition Technology : A Comprehensive Overview,” Academic Journal of Computing & Information Science, vol. 6, no. 7, pp. 15–26, 2023, doi: 10.25236/AJCIS.2023.060703.
A. Madan, “Face Recognition using Haar Cascade Classifier,” International Journal for Modern Trends in Science and Technology, vol. 7, no. 01, pp. 85–87, 2021.
A. H. Ahmad et al., “Real time face recognition of video surveillance system using haar cascade classifier,” Indones. J. Electr. Eng. Comput. Sci., vol. 21, no. 3, pp. 1389–1399, Mar. 2021, doi: 10.11591/ijeecs.v21.i3.pp1389-1399.
P. Mccullagh, “Face detection by using Haar cascade classifier,” Wasit Journal of Computer and Mathematics Science, vol. 2, no. 1, pp. 1–5, 2023, doi: 10.31185/wjcm.109.
A. A. M. Suradi, I. Djafar, S. Alam, and A. Syam, “Perbandingan metode Haar cascade dan dlib dalam mendeteksi wajah secara realtime,” in Prosiding Seminar Ilmiah Sistem Informasi dan Teknologi Informasi, Makassar, Indonesia, pp. 94–102, 2023.
S. Saragih, “Implementasi Algoritma Haar Cascade Menggunakan Pengolahan Citra Digital untuk Absensi Deteksi Wajah dan Nama Menggunakan Python,” Jurnal Sosial dan Teknologi (SOSTECH), vol. 5, no. 3, pp. 789–798, 2025.
J. H. Sim and A. Yulianto, “Evaluating YOLOv5 and YOLOv8 : Advancements in Human Detection,” Journal of Information Systems and Informatics, vol. 6, no. 4, pp. 2999–3012, 2024, doi: 10.51519/journalisi.v6i4.944.
A. P. Sari and B. Hendrik, “Analisis komparatif algoritma deep learning untuk pengenalan wajah: CNN, FaceNet, dan ArcFace,” Journal of Education Research, vol. 6, no. 4, pp. 1029–1036, 2025.
B. W. Nurlita, S. Winarno, A. Nugraha, A. N. I. Muttaqin, Y. Zarifa, P. M. Salsabila, and G. F. Mumtaz, “Comparison of ArcFace and Dlib performance in face recognition with detection using YOLOv8,” Jurnal Inovtek Polbeng - Seri Informatika, vol. 9, no. 2, pp. 890–903, 2024.
T. V. Dang, “Smart home management system with face recognition based on ArcFace model in deep convolutional neural network,” Journal of Robotics and Control (JRC), vol. 3, no. 6, pp. 754–761, 2022, doi: 10.18196/jrc.v3i6.15978.
S. Kishore, P. H. G., S. C. S., and M. K. B., “Evaluation of deep learning methods in face recognition: Datasets, metrics, and results,” International Journal on Science and Technology (IJSAT), vol. 16, no. 4, pp. 1–9, 2025.
D. S. Alfan, A. Rochman, M. Firdaus, N. Setiawan, and P. Rosyani, “Penerapan metode Haar-cascade dan LBPH untuk face detection dan recognition,” Jurnal AI dan SPK: Jurnal Artificial Inteligent dan Sistem Penunjang Keputusan, vol. 2, no. 1, pp. 96–104, 2024.
F. L. Ramadini and E. Haryatmi, “Penggunaan metode Haar cascade classifier dan LBPH untuk pengenalan wajah secara realtime,” InfoTekJar: Jurnal Nasional Informatika dan Teknologi Jaringan, vol. 6, no. 2, pp. 289–296, 2022, doi: 10.30743/infotekjar.v6i2.4714.
E. Abdullahu, H. Wache, and M. Piangerelli, “Secure and decentralized hybrid multi-face recognition for IoT applications,” Sensors, vol. 25, p. 5880, 2025, doi: 10.3390/s25185880.
Raspberry Pi Ltd, “Raspberry Pi 5 Product Brief,” Dec. 2025. [Online]. Available: https://pip-assets.raspberrypi.com/categories/892-raspberry-pi-5/documents/RP-008348-DS-6-raspberry-pi-5-product-brief.pdf. [Accessed: Apr. 13, 2026].
S. M. H. Abidi, S. A. Hassan, S. M. Raza, and M. J. Beliatis, “Advances in face recognition: A comprehensive review of feature extraction and dataset evaluation,” Electronics, vol. 15, p. 338, 2026, doi: 10.3390/electronics15020338.
How To Cite This :
Refbacks
- There are currently no refbacks.
