Date of Publication :15th March 2025

Abstract: Skin diseases are one of the most prevalent health concerns in canines, with an overall dermatological disorder prevalence of 33.4%. Early and accurate detection is crucial for effective treatment and management. This study proposes a Convolutional Neural Network (CNN)-based deep learning model to classify and predict a diverse range of skin diseases in dogs using image-based analysis. Unlike existing approaches that primarily focus on bacterial infections, fungal i nfections, and hypersensitivity-related allergies, our model expands its scope to include viral infections, tumors, and abnormal skin growths, offering a comprehensive diagnostic tool. The model is trained on a curated dataset with data augmentation techni ques to improve generalization and tested using InceptionV3, a state- of-the-art CNN architecture, ensuring high accuracy and robustness. The system is deployed as a user-friendly web interface, allowing pet owners and veterinarians to upload images for re al- time disease prediction along with confidence scores. By automating the diagnosis of multiple canine skin conditions, this research aims to reduce manual veterinary workload, improve early detection rates, and enhance pet healthcare. The findings demonstrate that deep learning-based models can significantly improve diagnostic accuracy and accessibility, contributing to better veterinary care and disease management practices.

Reference :