CLASSIFICATION OF SIBERIAN HUSKY AND GOLDEN RETRIEVER DOGS USING CONVOLUTIONAL NEURAL NETWORK METHOD

Abstract

Classification of a dog is interesting because it allows owners to provide specialised care, detect health issues, and track relevant genetic history, ultimately improving the well-being and lives of dogs. Dog classification based on images is an interesting and crucial task in the fields of pattern recognition and computer vision. Convolutional Neural Network (CNN) methods have become one of the most effective and popular approaches for classifying objects in images. CNN is a type of artificial neural network architecture inspired by the way human vision works. In this research, the author has developed a Python program using CNN to recognize Golden Retrievers and Huskies. The program is designed to classify these specific dog breeds based on provided images. The research is conducted through several stages to achieve the desired goal. These stages include data collection, data preprocessing, CNN architecture design, model training, model evaluation, validation and testing, as well as result analysis. Based on the executed program, the CNN method successfully classifies Huskies and Golden Retrievers effectively. The accuracy that is displayed by the graphs shows an improvement of accuracy from time to time. The model loss also decreases from time to time. In conclusion, the CNN method can achieve high accuracy levels up to 100%  in object classification.