A Deep Research in Classic Classification Network

Abstract

Abstract This article is a further study and reflection of the classification network after the completion of the article “The Comparison and Analysis of Classic Convolutional Neural Network in the Field of Computer Vision”. Different from previous work comparing network performance of VGGNet and ResNet, this paper compares and analyzes the importance of depth, breadth and residual learning of GoogLeNet and ResNet in convolutional neural networks. And considering that most of the time, the basic classification network nowadays appears in the form of backbone in the detection and segmentation network, this paper compares the advantages and disadvantages of several major classification networks as backbones.