An optimized CNN–LSTM model for analyzing data from surveillance cameras

Abstract

This paper presents an artificial intelligence–based approach for analyzing educational processes using data obtained from surveillance cameras. The increasing deployment of surveillance cameras in urban, industrial, and transportation environments has led to the generation of massive volumes of video data, creating significant challenges for automated monitoring and real-time analysis. Traditional computer vision methods often struggle with variations in lighting, occlusions, and dynamic scenes, resulting in reduced accuracy and efficiency. This study proposes an optimized Convolutional Neural Network–Long Short-Term Memory (CNN–LSTM) model designed to address these limitations by simultaneously extracting spatial and temporal features from surveillance video streams. The proposed model is evaluated on benchmark surveillance datasets, demonstrating superior performance in object detection, activity recognition, and anomaly detection compared to conventional CNN, LSTM, and baseline hybrid approaches. Experimental results indicate that the optimized CNN–LSTM model achieves higher precision, recall, and F1-scores while maintaining real-time processing capabilities, making it suitable for practical deployment in large-scale surveillance systems. The study provides a foundation for further research in real-time surveillance analytics and intelligent video monitoring systems.