An Optimized Object Detection System using You Only Look Once Algorithm and Compare with Deep Neural Networks with increased

Abstract

The main objective of this research is to detect and track and recognize objects of different sizes using the novel you only look once algorithm and compare its accuracy with deep neural networks. Various kinds of objects are detected, recognized and tracked using the Novel You Only Look Once algorithm ($\mathrm{N}=10$) and its accuracy is compared with deep neural networks ($\mathrm{N}=10$) algorithm. In the research, the proposed algorithm i.e, the Novel You Only Look Once algorithm and Deep Neural networks techniques are utilized and the accuracy between those techniques are analysed. The accuracy of Deep Neural Networks is about 89.18% and using the Novel You Only Look Once algorithm is 97.48% with ($\mathrm{p} < 0.005$). Upon researching, studying various thesis and books and inferring results, it was proved that the novel you only look once algorithm has a greater accuracy than deep neural networks.