Comparative Analysis of Object Detection Algorithms

Abstract

The field of image processing has attracted a lot of attention during the last decade. Object detection algorithms have seen rapid development from conventional architectures to more sophisticated architectures which rely on the neural networks for cognitive pattern recognition. Sophisticated machine learning algorithms and faster GPUs have rendered us with a plethora of algorithms for object classification as well as object detection, the most prominent ones have been discussed in this report. Our main objective is to compare object classification and object detection models. From the number of proposed models over the years, this work picks the best, “state of the art” object detection models for comparison, namely You Only Look Once and Single Shot Multibox Detector. Moreover, this work also compares the underlying backbone architecture of these models and how well they fare off against each other.