Object Detection Based on Deep Learning: A Brief Review

Abstract

Object Detection is one of the basic tasks in the field of computer vision. In recent years, with the hot development of deep learning technology, object detection algorithm has changed from the traditional algorithm based on manual features to the detection technology based on deep neural network. From the first proposed R-CNN, OverFeat, to the later Fast/Faster RCNN, SSD, YOLO series, and most recently Pelee. In less than five years, the object detection technology based on deep learning has improved the network structure from two stages to one stage, from bottom-up only to Top-Down. From single-scale network to feature pyramid network, from PC side to mobile side, many good algorithm technologies have emerged, and these algorithms have excellent detection effect and performance on open target detection data sets. This paper first reviews the traditional target detection algorithms, then introduces several mainstream two-stage target detection algorithms and single-stage target detection algorithms, analyzes the structure, advantages and disadvantages of these algorithms, and finally forecasts the future research direction of target detection algorithms.