International Journal of Sustainable Building Technology and Urban Development ISSN:2093-761X(Print) 2093-7628(Online)

Sustainable building environments increasingly rely on intelligent monitoring systems to enhance safety, optimize facility management, and support secure urban infrastructure. This study presents an intelligent video monitoring framework for behavioral analysis and risk assessment within sustainable building environments using classical machine learning techniques. The proposed non-invasive system analyzes visual cues derived from facial and hand gestures captured through surveillance video streams. Nine behavioral features were extracted and categorized into facial and hand gesture modalities. Seven supervised classification algorithms—Logistic Regression, Decision Tree, k-Nearest Neighbor (KNN), Linear Discriminant Analysis (LDA), Gaussian Naïve Bayes, Support Vector Machine (SVM), and Random Forest—were implemented to evaluate system performance. Comparative analysis demonstrates that KNN and SVM achieved the highest overall accuracy of approximately 80% when multimodal gesture features were fused. Individually, facial gestures yielded up to 79% accuracy using LDA, while hand gestures achieved 72% accuracy. The results indicate that classical machine learning methods can effectively support behavioral risk detection in smart building environments. The proposed framework offers a cost-effective and scalable solution that can be integrated into sustainable building management systems to enhance safety monitoring and contribute to resilient urban infrastructure.