Human activity recognition from accelerometer data using Convolutional Neural Network
Song-Mi Lee, Sang Min Yoon, Heeryon Cho · 2017
We propose a one-dimensional (1D) Convolutional Neural Network (CNN)-based method for recognizing human activity using triaxial accelerometer data collected from users' smartphones. The three human activity data, walking, running, and staying still, are gathered using smartphone accelerometer sensor. The x, y, and z acceleration data are transformed into a vector magnitude data and used as the input for learning the 1D CNN. The ternary activity recognition performance of our 1D CNN-based method which showed 92.71% accuracy outperformed the baseline random forest approach of 89.10%.