Animal face classification based on deep learning

Using deep learning to automatically identify and classify wildlife can greatly improve the efficiency of wildlife protection strategies. Based on the public data set of Animal Faces-HQ, this essay analyzed methods to classify different types of animals. Then, this essay developed the four means to predict animal species. First two adopts traditional Convolutional neural networks, while later two construct the deep-learning model based on Auto Encoder. Experimental results showed that our four methods is accuracy and effectiveness while the model in which SVM is superimposed on neural network feature extractor performs slightly better. To verify the applicability of our model, this research applied this model in additional test set of data, and it also achieved satisfactory fitting results.