Review of Target Detection Technology based on Deep Learning

Abstract

Target detection is one of the most important contents in computer vision, which has been widely and effectively applied in production, daily life, and military. The target detection technology based on deep learning has gone far beyond the traditional target detection technology with high autonomy, accuracy and sensitivity. Based on convolutional neural network, it mainly develops two-stage detection algorithm and one-stage detection algorithm. The two-stage detection algorithms mainly include RCNN, SPP-Net, Fast-RCNN, Faster-RCNN, etc., and the one-stage detection algorithms mainly include Yolo, SSD, RetinaNet, etc. Deep learning framework is an important tool to implements target detection algorithm. Current mainstream deep learning frameworks include PyTorch, TensorFlow, PaddlePaddle, keras, etc.