| Title |
Deep Learning-based Telemedicine Platform for Diagnosing Skin Diseases in Pet Dog |
| Authors |
류가연(Gayeon Ryu) ; 최경택(Kyoungtaek Choi) |
| DOI |
https://doi.org/10.5573/ieie.2026.63.2.96 |
| Keywords |
Pet; Skin diseases; Transfer learning; Deep learning; Lesion-specific classifier |
| Abstract |
In this study, we propose a remote diagnostic system that predicts skin diseases in pet dogs by analyzing images captured with a mobile phone camera. Users manually specify the affected area, and the system then performs disease classification based on the selected region. We validated the system using a large-scale canine dermatology dataset provided by AI-HUB. Considering the common issue of label noise in medical datasets, we conducted comparative experiments using both the original (noisy) data and a refined version. Furthermore, we compared the performance of a single unified classifier with that of separate classifiers trained for each specific body region. The region-specific classifiers achieved a classification accuracy of 82.19% across six different skin diseases, domonstrating the effectiveness of our approach. |