딥러닝 기반 사진 분석을 활용한 실종 반려견 찾기 시스템 설계와 구현

초록

This study proposes a method for designing and implementing a lost pet finding system using deep learning-based image analysis. Traditional methods for locating lost pets are time-consuming, costly, and have low success rates, leading to psychological burdens for pet owners. To address this issue, this study presents an integrated system that allows pet owners to upload photos of lost pets and receive real-time notifications. Users upload photos of the lost pet, and the system uses a deep learning model to classify whether the animal is a dog and identify the breed. The system is designed with a client-server architecture, integrating user authentication, data management, model calls, and notification features to provide real-time alerts. Experimental results show that the dog identification model achieved an accuracy of 99.3%, and the breed classification model achieved an accuracy of 93.5%. This study proposes a new approach to solving the problem of lost pets and is expandable to address the issues of missing other types of pets in the future.

키워드