Training Classifiers with Shadow Features for Sensor-Based Human Activity Recognition.

Simon Fong,Kelvin Wong,Wei Song,Kyungeun Cho,Raymond Wong

doi:10.3390/s17030476

Abstract

In this paper, a novel training/testing process for building/using a classification model based on human activity recognition (HAR) is proposed. Traditionally, HAR has been accomplished by a classifier that learns the activities of a person by training with skeletal data obtained from a motion sensor, such as Microsoft Kinect. These skeletal data are the spatial coordinates (x, y, z) of different parts of the human body. The numeric information forms time series, temporal records of movement sequences that can be used for training a classifier. In addition to the spatial features that describe current positions in the skeletal data, new features called ‘shadow features’ are used to improve the supervised learning efficacy of the classifier. Shadow features are inferred from the dynamics of body movements, and thereby modelling the underlying momentum of the performed activities. They provide extra dimensions of information for characterising activities in the classification process, and thereby significantly improve the classification accuracy. Two cases of HAR are tested using a classification model trained with shadow features: one is by using wearable sensor and the other is by a Kinect-based remote sensor. Our experiments can demonstrate the advantages of the new method, which will have an impact on human activity detection research.

Highlights

Human activity recognition (HAR) is a branch of machine learning that trains a classification model using historical activity records to recognise unseen activities
SensorsHowever, 2017, 17, 476for this paper, we argue that statistical features and wavelet features are not always the most appropriate for building classification models for HAR
The first collection of results done with accelerometer data and Kinect data are shown in butterfly charts in Figures 8a–c and 9a–c, respectively

Summary

Introduction

Human activity recognition (HAR) is a branch of machine learning that trains a classification model using historical activity records to recognise unseen activities. The sensor data arrive in the form of continuous and sequential values, similar to time series. Time series, such as those collected from an accelerometer, are usually multivariate, comprising tri-axial spatial information, known in its simplest form as ‘three features’ (x, y, z). These three features, together with the time-stamp information in the time series, represent the temporal-spatial displacement of a body part in motion.

Methods

Results

Conclusion