Author(s) :

Rajat.

Volume/Issue :

Abstract :

Real-time object detection is a much-needed subsection of computer vision and can be applied to autonomous vehicles, surveillance, medical imaging, and robotics. This research paper provides an overview of real-time object detection strategies, specifically within the architectures of real-time object detection methods such as SSD (Single Shot Detector), and with Faster R-CNN based on a deep learning architecture. This paper also discusses the advantages of these architectures, limitations, and performance metrics. The proposed approach discussed in this paper integrates a hybrid deep learning method optimized for low-latency accuracy in real-time detection. The Convolutional Neural Network (CNN) is essential to modern object detection networks, and is the backend of OpenCV, which is an open-source based computer vision library supporting computer-based vision applications. OpenCV library contains over 2,500 optimized algorithms to implement for tasks such as face recognition, object tracking. Its open-source nature makes it highly inexpensive and adaptable to a business or research setting, greatly enhancing the innovation of real-time computer vision applications.

No. of Downloads :

0