상세 보기
- 김벼리;
- 곽민재;
- 추현규;
- 이수민;
- 위재빈;
- ... 전주현;
- 외 1명
WEB OF SCIENCE
0SCOPUS
0초록
In this study, we developed three models to classify R, G, B colors that are captured differently by the camera depending on the surrounding lighting using machine learning and deep learning libraries, and conducted an experiment to classify data collected by Orange Pi. Three Classification models were implemented using a decision tree, random forest, and CNN algorithm. As a result of performance evaluation using ROC_AUC score, values of 99.6%, 98.6%, 99.9% were derived respectively. Also, when considering Accuracy, Precision, Recall, and F1 score comprehensively, it was confirmed that they showed better performance in the order of CNN algorithm, Decision Tree, and Random Forest. In addition, by successfully conducting experiments on classification using photo data collected through cost-effective Orange Pi, we have shown that AI-related projects can be carried out sufficiently with low-cost devices.
키워드
- 제목
- 오렌지 파이 하드웨어와 AI를 활용한 RGB 색상분류 모델 개발 및 성능분석
- 제목 (타언어)
- Development and performance analysis of RGB Color classification model using Orange Pi hardware and AI
- 저자
- 김벼리; 곽민재; 추현규; 이수민; 위재빈; 김나현; 전주현
- 발행일
- 2022-12
- 저널명
- 지식재산 교육과 연구
- 권
- 10
- 호
- 2
- 페이지
- 27 ~ 36