ScholarWorks@숙명여자대학교

초록

To date, the number of people who have companion animals has gradually increased and the need for advancement in veterinary care and pet health care has been increased. Deep learning models are taking their places in healthcare and can be used for detecting diseases. We aimed to build and validate a framework for auxiliary diagnosis of pet diseases in everyday life before hospital visits. Our framework utilizes disease image classification and natural language models with Swin-Transformer and Bidirectional Encoder Representations from Transformers as the backbone, respectively, and both presented the accuracy of 84.5% and 84%, respectively. This proposed framework can be useful in understanding animals’ symptoms for pet owners as well as assisting a veterinarian for diagnosis.

키워드