Author(s) :   Kajal Kharabe

Abstract : The importance of object detection in video analysis and picture interpretation has led to a recent spike in interest in the field. The performance of traditional object detection techniques was constrained by their reliance on manually created features and simple trainable algorithms. The inadequacies of earlier methods have been addressed, though, by the development of deep learning (DL), which has produced increasingly potent tools that can extract deep, high-level, and semantic information. There are differences in network design, training methods, and optimization functions amongst Deep Learning-based object identification models. This study looked into popular general designs for object detection as well as different adjustments and suggestions to improve detection performance. In addition, the problems and future prospects of object identification research have been highlighted, along with developments in neural network-based learning systems. Furthermore, a comparison study based on performance metrics of several iterations of the Yolo technique for multiple object identification has been provided.

Keywords: object identification, neural networks, deep learning, and YOLO.

DOI : 10.61161/ijarcsms.v12i7.42

Pages : 339-343

*Authors are invited to submit papers through E-mail at editor.ijarcsms@gmail.com