Learning Multi-level Features For Sensor-based Human Action Recognition

TL;DR

This paper introduces a multi-level feature learning framework for sensor-based human action recognition, combining signal, component, and semantic analysis to improve accuracy using a single inertial sensor.

Contribution

It presents a novel multi-level feature learning framework with the MLPL method for high-level semantic understanding, achieving state-of-the-art results.

Findings

Achieved 88.7% weighted F1 on Skoda dataset.

Achieved 98.8% weighted F1 on WISDM dataset.

Achieved 72.6% weighted F1 on OPP dataset.

Abstract

This paper proposes a multi-level feature learning framework for human action recognition using a single body-worn inertial sensor. The framework consists of three phases, respectively designed to analyze signal-based (low-level), components (mid-level) and semantic (high-level) information. Low-level features capture the time and frequency domain property while mid-level representations learn the composition of the action. The Max-margin Latent Pattern Learning (MLPL) method is proposed to learn high-level semantic descriptions of latent action patterns as the output of our framework. The proposed method achieves the state-of-the-art performances, 88.7%, 98.8% and 72.6% (weighted F1 score) respectively, on Skoda, WISDM and OPP datasets.