메뉴 건너뛰기
소속 기관 / 학교 인증
인증하면 논문, 학술자료 등을  무료로 열람할 수 있어요.
한국대학교, 누리자동차, 시립도서관 등 나의 기관을 확인해보세요
(국내 대학 90% 이상 구독 중)
고객센터 ENG
주제분류

논문 기본 정보

저자정보
(서울대학교) (서울대학교) (서울대학교) (서울대학교)
저널정보
한국HCI학회 한국HCI학회 학술대회 PROCEEDINGS OF HCI KOREA 2018 학술대회 발표 논문집
오류 신고하기

피인용 1

검색

    초록·키워드

    The aim of this study is to build human activity recognition (HAR) model using deep neural network (DNN) and investigate the influence that affects misclassification. As wearable devices become widespread and used in various applications such as health care and sports, people are interested in HAR. Therefore, it is important to improve classification performance in HAR. We implemented a DNN based HAR model through open smartphone sensor data set and t-Distributed Stochastic Neighbor Embedding was used to visualize extracted features. The performance of the DNN model was excellent except for one activity. Through the visualization of the extracted features, we were able to identify the cause of the performance degradation. Similar extracted features between activities are the cause of performance degradation. The DNN model can recognize human activity using smart phone sensor data and be used for health care, sports, fall detection and so on.

    최근 본 자료 전체보기

      UCI(KEPA) : I410-ECN-0101-2018-004-001761610