Animal Image Classifier Based on Convolutional Neural Network

In modern society, there are dogs and cats around people, as well rare wild animals living in nature. The relationship between human beings is getting closer closer. rapid development of machine learning deep technology has been widely used the academic field. Aiming at problem animal image classification, this paper uses Pytorch to learn about 10,000 pictures containing cats, dogs, (tiger, lion, etc.) based on research algorithm convolutional neural network field classification. And a model that can realize classifier established optimized, so efficiently classify wildlife pictures. results show accuracy two models above 90%, loss ranges from 0.706 0.061, 0.807 0.051, respectively, showing characteristics good fitting effect strong optimization ability. Meanwhile, be increased by properly increasing number full connection layers. Therefore, constructing network, accurate detection national ecological protection images realized.