Enhanced Object Detection and Distance Prediction for Autonomous Vehicles Using Deep Neural Network
Author : R Gomathi Jayam and R Shobana
Abstract :
The proposed system is an efficient real-time video processing framework designed for accurate object detection and visual analysis using advanced deep neural network techniques. It supports both live camera feeds and prerecorded videos, making it suitable for various real-world applications. The system preprocesses video frames through resizing, normalization, and blob conversion to ensure fast and stable performance under real-time conditions. A deep learning model detects objects with high precision, while confidence-based filtering removes unreliable predictions to improve accuracy. Additional features such as gesture analysis and object tracking enhance adaptability for specialized tasks. The framework includes a real-time performance monitor to maintain low latency and displays annotated video output for easy interpretation. Overall, it provides a scalable, high-accuracy solution for applications like intelligent surveillance, robotics, industrial automation, and autonomous navigation.
Keywords :
Real-Time Video Processing, Object Detection, Deep Neural Networks (DNN), Computer Vision, Frame Preprocessing, Confidence Filtering, Gesture Analysis, Object Tracking, Intelligent Surveillance, Autonomous Systems.
