Research on Target Detection algorithm based on Deep Learning Technology

This paper summarizes the research progress of target detection using convolution neural network in recent years. These studies not only cover the design of all kinds of convolution neural network target detection algorithms, but also provide a deeper perspective for the development of computer vision. On the basis of consulting the data, this paper focuses on the representative Faster-RCNN, YOLO V3 and SSD algorithms. By reviewing their predecessor algorithms, covering the current mainstream target detection algorithms, and analyzing the technologies they use, summarize and analyze their advantages and disadvantages. And in the last part, it points out the still existing problems in target detection and the development direction in the future.