AI Video Detection using Convolutional Neural Networks

Abstract

This project develops an Artificially Intelligent (AI) video detection and analysis tool by using Convolutional Neural Network (CNN). The tool is based on an effective Python architecture that uses OpenCV and specially trained YOLOv8 CNN models. Its three main features include object detection for security monitoring, age estimation with demographic analysis, and facial expression recognition for eight emotional states (Anger, Contempt, Disgust, Fear, Happy, Neutral, Sad, and Surprise). The tool offers hands-on experience with computer vision techniques and security surveillance through real-time video processing with overlay annotations and analytics dashboards. The trained model achieves 83.6% accuracy in facial expression detection with real-time processing capabilities for comprehensive video analysis. This CNN-based AI video detection tool was created as a learning platform for students pursuing computer vision and artificial intelligence. It combines theoretical understanding with hands-on application of convolutional neural network technologies for practical security, surveillance uses, and digital forensic image analysis for investigative purposes.