%0 Electronic Article
%A H. Cho
%A S.M. Yoon
%K CNN
%K 1D convolutional neural network
%K public falls
%K principal component analysis based acceleration features
%K singular value decomposition
%K activity recognition
%K triaxial accelerometer data
%K one-dimensional convolutional neural network based fall
%K fall detection
%K kernel principal component analysis
%K raw acceleration data
%K useful features
%K three-dimensional reduction methods
%K sparse principal component analysis
%K SVD
%X The usefulness of applying singular value decomposition (SVD) on triaxial accelerometer data for one-dimensional (1D) convolutional neural network (CNN) based fall and activity recognition is investigated. Three-dimensional reduction methods, namely, SVD, sparse principal component analysis, and kernel principal component analysis, are compared for their effectiveness in extracting useful features for fall and activity recognition. Experiments conducted on three public falls and activities of daily living datasets show that SVD applied to acceleration data coupled with raw acceleration data or acceleration signal magnitude vector exhibited better 1D CNN fall and activity recognition accuracy than those using other principal component analysis based acceleration features.
%@ 0013-5194
%T Applying singular value decomposition on accelerometer data for 1D convolutional neural network based fall detection
%B Electronics Letters
%D January 2019
%I Institution of Engineering and Technology
%U https://digital-library.theiet.org/;jsessionid=2fbps9rqt2pgg.x-iet-live-01content/journals/10.1049/el.2018.6117
%G EN