Human Activity Recognition from Accelerometer Data with Convolutional Neural Networks

Abstract

Smartphones are present in most people's daily lives. Sensors embedded in these devices open the possibility of monitoring users’ activities. The classification of the intricate data patterns collected through these sensors is a challenging task when considering hand-crafted features and pattern recognition algorithms. In this work, to face this challenge, we propose a convolutional neural network architecture along with two methods for transforming sensor data stream into images. The proposed model was evaluated using the UniMiB SHAR dataset. The best macro average accuracy obtained for classification of 17 types of activities, with fivefold-cross-validation-method, was 90.44%.