• Indu bala
• Dr. Sunita Mahajan
• Dr. Ikvinderpal Singh

Real-time object detection plays a crucial role in various domains, including autonomous systems, surveillance, and robotics. Deep learning techniques have revolutionized object detection by providing state-of-the-art accuracy and speed. This research presents a comprehensive comparative study of deep learning architectures for real-time object detection. The study focuses on three widely-used architectures: Faster R-CNN, YOLO (You Only Look Once), and SSD (Single Shot MultiBox Detector).A diverse and annotated dataset was used for training and evaluation. The dataset preprocessing involved augmentation and normalization. The training process encompassed hyperparameter tuning and optimization. The evaluation metrics included precision, recall, F1 score, and mean Average Precision (mAP). Additionally, the inference speed of each architecture was measured.The experimental results reveal nuanced trade-offs between accuracy and speed. Faster R-CNN demonstrated exceptional accuracy with slightly lower inference speed, making it suitable for applications prioritizing precision. YOLO exhibited competitive accuracy with a notable increase in speed, positioning it as a strong choice for real-time scenarios. SSD demonstrated a balanced trade-off between accuracy and speed. This comparative study sheds light on the strengths and weaknesses of different deep learning architectures for real-time object detection. The findings provide valuable insights for selecting the most appropriate architecture based on application requirements. Future research directions include exploring hybrid architectures and optimizing trade-offs further to meet evolving real-world demands.

Object Detection; Faster R-CNN; YOLO; SSD; COCO Dataset;

"Detection of Real-Time Objects using Deep Learning Techniques", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.11, Issue 1, page no.b747-b752, January-2024, Available :http://www.jetir.org/papers/JETIR2401187.pdf

"Detection of Real-Time Objects using Deep Learning Techniques", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.11, Issue 1, page no. ppb747-b752, January-2024, Available at : http://www.jetir.org/papers/JETIR2401187.pdf