Integrating Object Detection and Deep Convolutional Neural Networks for Cat Breed Classification

Yiming Feng

2024

Abstract

This study presents a novel approach to cat breed identification using a combination of object detection and deep learning classification models. The project's primary objective is to assist animal shelters and adoption centers by accurately identifying the breeds of homeless cats from images, thereby enhancing the cats' chances of finding suitable homes and facilitating targeted healthcare. This work utilized the convolutional neural network model, pre-trained on ImageNet, and integrated it with OpenCV for initial cat detection. The dataset, comprising five major cat breeds—Calico, Persian, Siamese, Tortoiseshell, and Tuxedo—was subjected to rigorous preprocessing, including image and metadata matching, cat detection, data augmentation, and rescaling. The model was trained and tested for breed classification, achieving an impressive accuracy of 87%. The integration of detection and classification not only improved the model's focus on relevant image features but also bolstered its robustness against background noise and variations in image quality. The findings underscore the potential of deep learning in animal breed identification, offering a scalable solution for broader applications in animal care and research.

Paper Citation

in Harvard Style

Feng Y. (2024). Integrating Object Detection and Deep Convolutional Neural Networks for Cat Breed Classification. In Proceedings of the 1st International Conference on Modern Logistics and Supply Chain Management - Volume 1: MLSCM; ISBN 978-989-758-738-2, SciTePress, pages 433-438. DOI: 10.5220/0013337500004558

in Bibtex Style

@conference{mlscm24,
author={Yiming Feng},
title={Integrating Object Detection and Deep Convolutional Neural Networks for Cat Breed Classification},
booktitle={Proceedings of the 1st International Conference on Modern Logistics and Supply Chain Management - Volume 1: MLSCM},
year={2024},
pages={433-438},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013337500004558},
isbn={978-989-758-738-2},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 1st International Conference on Modern Logistics and Supply Chain Management - Volume 1: MLSCM
TI - Integrating Object Detection and Deep Convolutional Neural Networks for Cat Breed Classification
SN - 978-989-758-738-2
AU - Feng Y.
PY - 2024
SP - 433
EP - 438
DO - 10.5220/0013337500004558
PB - SciTePress