Image Recognition of Different Hamster Breeds Using Convolutional Neural Networks

Hamsters exhibit a variety of breeds, making breed identification challenging at first glance. This study proposes a solution using Convolutional Neural Networks (CNNs), specifically VGG16 architecture, to aid pet owners in distinguishing between hamster breeds. Researchers developed an image capturing system and a web application where users can upload hamster images to receive breed identification. The system focuses on recognizing the four most common hamster breeds: Campbell, Roborovski, Syrian, and Winter White, utilizing the VGG16 model trained and tested on a dataset collected by the researchers. The implementation involved a Raspberry Pi 4B with 4GB RAM and a webcam housed in a specially designed box for image recognition of hamsters. Evaluation using a confusion matrix demonstrated an overall accuracy of 91.67%, indicating the effectiveness of the proposed system in identifying hamster breeds. This research contributes to the development of accessible tools for pet owners and enthusiasts interested in hamster breed identification.