BCG Artifact Removal Using Improved Independent Component Analysis Approach

Abstract

Recent advancement in bio-medical field has attracted researchers toward BCG signal processing for monitoring the health activities. There have been various techniques for monitoring physical activities such as (SCG) Seismocardiography, Electrocardiography (ECG) etc. BCG signal is a measurement of reaction force applied for cardiac ejection of blood. Various measurement schemes and systems have been developed for BCG detection and measurement such as tables, beds, weighing scale and chairs. Weighing scales have been promising method for measurement of BCG signal because of less cost of implementation, smaller size etc. but these devices still suffer from the artifact which are induced due to subject movement or motion during signal acquisition or it can be caused due to floor vibrations. Artifact removal is necessary for efficient analysis and health monitoring. In this work we address the issue of artifact removal in BCG signal by proposing a novel method of signal processing. According to proposed approach raw signal is pre-processed and parsed to independent component analysis which provides the decomposed components and later k-means is applied to detect the components which are responsible for artifact and removed. Proposed approach is compared with existing method and shows better performance in terms of artifact removal.