COMPARE SSD ALGORITHM AND YOLO IN OBJECT DETECTION

Abstract

Object detection can be divided into two groups: 1) detecting a specific object, and 2) detecting categories of the object. Most of methods based on R-CNN family (Regions with Convolutional Neural Network Family) such as R-CNN, Fast R-CNN, Faster R-CNN, Mask R-CNN,… it comprises of a sequence of processing of alternated layers which is very complex and expensive. In 2016, Joseph Redmon et al. proposed a method of object detection named YOLO (You Only Look Once), and Wei Liu et al. proposed a method of object detection named SDD (Single Shot Detector) based on a different approach