Convolutional Neural Network Based Dog Breed Classification

Abstract

One of the most popular domestic animals is the dog. Numerous problems, including population control, a decline in disease outbreaks like rabies, vaccination oversight, and legal ownership, are brought on by the abundance of dogs. There are currently around 180 different dog breeds. Each breed of dog has unique traits and health issues. It is crucial to identify people and their breeds in order to administer the proper therapies and training. Machine learning provides the ability to train algorithms that can tackle the challenges of information classification and prediction based only on newly emerging information as raw data. For the categorization and detection of images, Convolutional Neural Networks (CNNs) provide a single, widely utilized method. In this effort, we establish a CNN-based method for identifying dogs in potentially complicated photos, and as a result, we consider only one of the types of dog breed identification. The graphical depiction demonstrates that the algorithm (CNN) delivers good analysis accuracy for all the examined datasets because the experimental outcome analysis confirmed the conventional metrics.