Open AccessAnalysis of the Topological Structure of the Convolution Neural Network Model RESNET
Author(s) -
Tao Shang,
Ying Song
Publication year - 2020
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1575/1/012136
Subject(s) - convolutional neural network , convolution (computer science) , computer science , artificial neural network , pattern recognition (psychology) , artificial intelligence , residual neural network , network structure , image (mathematics) , deep learning , topology (electrical circuits) , field (mathematics) , network model , time delay neural network , algorithm , machine learning , mathematics , combinatorics , pure mathematics
With the advent of the era of big data and the development of artificial neural network, a large number of different neural network models have come out. In the field of image recognition, the application proportion of convolutional neural network model is very high, and the different structure levels of the same convolutional neural network model are also very different, and the recognition effect in solving different problems is not the same. Taking the application of image recognition as an example, this paper analyzes and compares the influence of different topological structures in RESNET model on recognition accuracy. Through the experimental comparison and analysis, it is concluded that the structure of shortcut_6 in RESNET convolutional neural network is the turning point. When the length of shortcut is less than 6, the recognition efficiency is on the rising trend with the deepening of the network. When it is equal to 6, the recognition efficiency is basically fixed, and when it exceeds 6, the recognition accuracy is reduced.