http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
점진적 학습법과 준 지도 학습법을 활용한 센서 기반 인간 행동 인식 프레임워크
안성호(Seong-Ho Ahn),전설(Seol Jeon),정동화(Dong-Hwa Jeong) 대한전자공학회 2022 대한전자공학회 학술대회 Vol.2022 No.11
Sensor-based human activity recognition using machine learning methods has been used in various healthcare services recently, and their significance is increasing. However, previous methods have limitations with leakage of personal information, increased computation time caused by data growth, and difficulty in data labeling. In this study, we propose a human activity recognition framework that combines incremental learning and semi-supervised learning methods to overcome these limitations. Our proposed method can be trained more quickly without performance loss than conventional batch learning methods in sensor-based human activity recognition.