상세 보기
초록
Among many dangerous situations, the number of cases of violence has been growing recently. However, there is currently no research to recognize conditions such as assault. Therefore, this paper presents a VR (Violence-Recognition) model for recognition activity using LSTM. The VR model develops algorithms that can detect dangerous situations through processing and analysis of sensing data. Also, to improve accuracy by using the FFT algorithm for processing digital signals in combination with LSTM. ? 2021, Springer Nature Singapore Pte Ltd.
키워드
Abnormal detection; Fusion sensing; LSTM; Smartphone; Smartwatch; Ubiquitous computing; Activity recognition; Dangerous situations; Digital signals; FFT algorithm; Sensing data; Long short-term memory
- 제목
- Activity-Recognition Model for Violence Behavior Using LSTM
- 저자
- Kim, Svetlana; Nam, Hyejeong; Park, Hyunho; Lee, Yong-Tae; Yoon, Yongik
- 발행일
- 2021-01
- 유형
- Conference Paper
- 권
- 715
- 페이지
- 529 ~ 535