Cat Breed Classification Using Kaggle Dataset Metadata with YOLO v5 Framework

Abstract

At this time cats have a variety of different breeds around the world including the Persian, Maine Coon, Siamese, Ragdoll, Sphynx and others. To find out, each cat breed can be seen from the pattern, coat color, and there are some faces that are different from other cats, but not completely the pattern, coat color and face can distinguish each cat breed. With the development of the times and increasing technology in the field of Computer Vision where the Artificial Intelligence system that is trained is used as a tool to classify types of cat breeds based on their faces using a computer. This study aims to be able to recognize and classify cat breeds based on their faces using YOLOv5. The evaluation parameters used are Confusion Matrix, Mean Average Precision, Precision and Recall. The experimental results show that the best model is achieved in the 60th epoch scenario in the 16th batch size with precision 0.9844, Recall of 1.0, mAP 0.5 of 0.9933 and mAP 0.5 : 0.95 of 0.9144.