Cat Breeds Classification Using Compound Model Scaling Convolutional Neural Networks.

Abstract

Cats are one of the most popular animals in the world. Many cat breeds in the world are only about 1%. Therefore, most are dominated by mixed cats or domestic cats. Nevertheless, there are so many different types of cat breeds in the world that it is sometimes difficult to identify them. Therefore, we need a system that can recognize and classify the types of cat breeds automatically. In this study, we used one of the deep learning methods that can recognize and classify an object, a Convolutional Neural Networks (CNN). The EfficientNet-B0 architecture was used as a model to extract image features automatically. The collection of nine different cat breeds containing 2700 images was used as a working dataset fed into the EfficientNet-B0 architecture. Based on the experiments, the system succeeds in classify cat breeds images, and the best model has achieved classification accuracy of 95%.