Animals Image Classification Method Based on Improved Convolutional Neural Network
Zihang Song, Wenkang Gong, Chenxi Li, Teoh Teik Toe · 2022 IEEE 10th Joint International Information Technology and Artificial Intelligence Conference (ITAIC) · 2022
This paper attempts to apply neural networks to classify images of 10 different animals. The dataset of the classification model is derived from tens of thousands of animal maps on the network, and these images have been artificially checked in advance. Convolutional neural network is a very popular way of image classification, is an important part of iconography. This article focuses on the analysis of the influence of the parameters of the algorithm we use on the model in order to optimize the model. Through the analysis of the network structure, we have improved the model using different methods to achieve a better model structure and application to classification. The final training accuracy is over 85% and the text precision is approximately 75%.