Claim Missing Document
Check
Articles

Real-Time Access Control System with YOLOv11-Based Face and Blink Detection Namira Nur Rifani; RD. Kusumanto; Nyayu Latifah Husni
Indonesian Journal of Artificial Intelligence and Data Mining Vol. 8 No. 3 (2025): November 2025
Publisher : Universitas Islam Negeri Sultan Syarif Kasim Riau

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

This study presents a real-time smart access control system that combines facial recognition with blink-based liveness detection to strengthen security and reduce spoofing risks. The main purpose is to provide a lightweight and efficient method that verifies both identity and physical presence in real time. The system employs two YOLOv11 models: one for detecting facial regions and another for distinguishing eye states through “open” and “closed” transitions. Identity verification is carried out by comparing facial embeddings using Euclidean distance. A private dataset was collected for facial images, while blink data was obtained from a public source, both annotated in YOLO format. After 100 epochs, the face detection model achieved 0.999 precision, 1.000 recall, 0.995 mAP50, and 0.868 mAP50–90, while the blink detection model recorded 0.959 precision, 0.962 recall, 0.967 mAP50, and 0.678 mAP50–90. These outcomes confirm that the objectives were achieved, demonstrating a practical and reliable biometric authentication solution with integrated liveness verification. The system also offers scalability for future multi-modal applications.
Comparative Evaluation of Optuna-Optimized Radial Basis Function and Sigmoid Kernels in Support Vector Machine for Smart Air Quality Classification Nanda Galea; A Rahman; Renny Maulidda; Nyayu Latifah Husni
Indonesian Journal of Artificial Intelligence and Data Mining Vol. 8 No. 3 (2025): November 2025
Publisher : Universitas Islam Negeri Sultan Syarif Kasim Riau

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

Poor air quality can have a serious impact on human health, so a classification system capable of accurately identifying air conditions is needed. This research proposes an air quality classification method using the Support Vector Machine (SVM) algorithm with two types of non-linear kernels, namely Radial Basis Function (RBF) and Sigmoid. The data used is obtained from various environmental sensors that record parameters such as CO, smoke, HC, TVOC, eCO₂, temperature, and humidity, and then collected in the form of historical datasets. To enhance the accuracy and efficiency of the model, hyperparameter optimization was performed automatically using Optuna. The evaluation results showed that SVM with RBF kernel performed better than Sigmoid kernel, achieving an accuracy value of 96.67% and F1-score of 96.80%. In addition, RBF also showed higher stability in 5-fold cross validation. This research shows that the combination of SVM and Optuna is effective in building an accurate air quality classification system, and has the potential to be further developed as a sensor based in air monitoring system and IoT.
Deteksi Aktivitas Buang Sampah Pada Video CCTV Menggunakan MoViNet-A0 Dengan Deteksi Manusia Berbasis YOLOv11 Raden Ayu Citra Dewi; R. D. Kusumanto; Nyayu Latifah Husni
Jurnal Informatika dan Teknik Elektro Terapan Vol. 14 No. 3 (2026)
Publisher : Universitas Lampung

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.23960/jitet.v14i3.10628

Abstract

Pembuangan sampah sembarangan di badan air perkotaan seperti Sungai Sekanak, Palembang, masih sulit diawasi secara manual melalui rekaman CCTV konvensional. Penelitian ini mengusulkan sistem deteksi aktivitas membuang sampah yang menggabungkan YOLOv11 sebagai detektor manusia untuk mengekstraksi Region of Interest (ROI) dan MoViNet-A0 sebagai model pengenalan aksi video untuk mengklasifikasikan aktivitas menjadi Buang Sampah dan Normal. Dataset penelitian terdiri dari 1.413 klip video CCTV (548 Buang Sampah dan 865 Normal) yang direkam pada siang hari di kawasan Sungai Sekanak dan dibagi dengan rasio 70:15:15 untuk training, validation, dan testing. Hasil pengujian pada data uji menunjukkan model mencapai accuracy sebesar 92,92%, dengan precision 92,41%, recall 89,02%, dan F1-score 90,68% pada kelas Buang Sampah, serta precision 93,23%, recall 95,38%, dan F1-score 94,30% pada kelas Normal. Pengujian real-time langsung menggunakan kamera CCTV pada 10 sampel aktivitas menghasilkan akurasi 70%, lebih rendah dibandingkan akurasi pelatihan, dengan kesalahan klasifikasi yang umumnya terjadi pada objek berjarak jauh dari kamera dan tertutup sebagian oleh vegetasi. Hasil ini menunjukkan bahwa kombinasi YOLOv11 dan MoViNet-A0 mampu mendeteksi aktivitas membuang sampah dengan akurasi tinggi pada data terkurasi, namun memerlukan penyempurnaan lebih lanjut agar lebih andal pada kondisi pengawasan nyata yang lebih bervariasi.