Dynamic Detection: Harnessing AI and Deep Learning for Object Detection in Moving Vehicles
- Version
- Download 38
- File Size 850.17 KB
- File Count 1
- Create Date 6 June 2024
- Last Updated 6 June 2024
Dynamic Detection: Harnessing AI and Deep Learning for Object Detection in Moving Vehicles
Dr. U.NILABAR NISHA1 , GOKULAKKANNAN A2 , JAYACHANDRAN J3,
PANIMAYA JERALD R4 , SELVAVINAYAGAM D5
1Head of the Department, Computer Science And Engineering, Mahendra Institute Of Engineering And Technology, Namakkal-637503
2Student, Computer Science And Engineering, Mahendra Institute Of Engineering And Technology, Namakkal-637503
3Student, Computer Science And Engineering, Mahendra Institute Of Engineering And Technology, Namakkal-637503
4Student, Computer Science And Engineering, Mahendra Institute Of Engineering And Technology, Namakkal-637503
5Student, Computer Science And Engineering, Mahendra Institute Of Engineering And Technology, Namakkal-637503
ABSTRACT- This paper introduce an AI-based approach utilizing deep learning for the detection of moving vehicles in video streams. Leveraging convolutional neural networks (CNNs) and recurrent neural networks (RNNs), our method achieves robust detection performance across diverse environmental conditions. We propose a multi-modal fusion strategy to integrate spatial and temporal information for improved accuracy. Experimental results on benchmark datasets demonstrate the effectiveness of our approach, outperforming existing methods with a detection accuracy exceeding 90% and real-time processing capabilities.we present a comprehensive framework for real-time moving vehicle detection using advanced AI and deep learning techniques. Our approach combines the power of convolutional neural networks (CNNs) for feature extraction and recurrent neural networks (RNNs) for capturing temporal dependencies in video sequences. We introduce a novel attention mechanism to focus on relevant spatial and temporal features, enhancing detection accuracy while reducing computational complexity. Through extensive experiments on challenging datasets, we demonstrate the superior performance of our method, achieving state-of-the-art results with over 95% accuracy and efficient processing rates suitable for real-world applications."
Download