Conference Paper

Detection of Atrial Fibrillation using model-based ECG analysis

Centre for Inf. & Syst., Univ. of Coimbra, Coimbra
DOI: 10.1109/ICPR.2008.4761755 Conference: Pattern Recognition, 2008. ICPR 2008. 19th International Conference on
Source: IEEE Xplore

ABSTRACT Atrial fibrillation (AF) is an arrhythmia that can lead to several patient risks. This kind of arrhythmia affects mostly elderly people, in particular those who suffer from heart failure (one of the main causes of hospitalization). Thus, detection of AF becomes decisive in the prevention of cardiac threats. In this paper an algorithm for AF detection based on a novel algorithm architecture and feature extraction methods is proposed. The aforementioned architecture is based on the analysis of the three main physiological characteristics of AF: i) P wave absence ii) heart rate irregularity and iii) atrial activity (AA). Discriminative features are extracted using model-based statistic and frequency based approaches. Sensitivity and specificity results (respectively, 93.80% and 96.09% using the MIT-BIH AF database) show that the proposed algorithm is able to outperform state-of-the-art methods.

1 Bookmark
  • [Show abstract] [Hide abstract]
    ABSTRACT: In this study we propose a novel atrial activity-based method for atrial fibrillation (AF) identification that detects the absence of normal sinus rhythm (SR) P-waves from the surface ECG. The proposed algorithm extracts nine features from P-waves during SR and develops a statistical model to describe the distribution of the features. The Expectation-Maximization algorithm is applied to a training set to create a multivariate Gaussian Mixture Model (GMM) of the feature space. This model is used to identify P-wave absence (PWA) and, in turn, AF. An optional post-processing stage, which takes a majority vote of successive outputs, is applied to improve classier performance. The algorithm was tested on 20 records in the MIT-BIH Atrial Fibrillation Database. Classification combining seven beats showed a sensitivity of 99.28%, a specificity of 90.21%. The presented algorithm has a classification performance comparable to current Heartrate-based algorithms yet is rate-independent and capable of making an AF determination in a few beats.
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: This study describes an atrial fibrillation (AF) detector whose structure is well-adapted both to detection of subclinical AF episodes and to implementation in a battery-powered device for use in continuous long-term monitoring applications. A key aspect for achieving these two properties is the use of an 8-beat sliding window, which thus is much shorter than the 128-beat window used in most existing AF detectors. The building blocks of the proposed detector include ectopic beat filtering, bigeminal suppression, characterization of RR interval irregularity, and signal fusion. With one design parameter, the performance can be tuned to put more emphasis on avoiding false alarms due to non-AF arrhythmias or more emphasis on detecting brief AF episodes. Despite its very simple structure, the results show that the detector performs better on the MIT–BIH Atrial Fibrillation database than do existing detectors, with high sensitivity and specificity (97.1% and 98.3%, respectively). The detector can be implemented with just a few arithmetical operations and does not require a large memory buffer thanks to the short window.
  • [Show abstract] [Hide abstract]
    ABSTRACT: In this study, we propose a P-wave absence (PWA) based method for atrial fibrillation (AF) identification over a short duration of electrocardiogram (ECG). The algorithm constructs a statistical model of normal sinus rhythm (SR) P-waves using a training set. Features extracted from P-waves are taken as an input to the Expectation–Maximization algorithm to create a Gaussian mixture model (GMM) of the P-wave feature space. The model is then used to identify PWA and detect AF. The algorithm performs AF identification in a single beat, and through post-processing of successive outputs using a majority voter determines the PWA over seven beats. The MIT-BIH Atrial Fibrillation Database was used to evaluate the algorithm. Classification using the majority voter showed a sensitivity of 98.09%, a specificity of 91.66%, a positive predictive value of 79.17% and an error of 6.88%. The performance of the proposed classifier is comparable to current R–R interval (RRI)-based algorithms, yet is able to detect short episodes of AF and performs rate-independent AF determination. The proposed algorithm targets atrial activity rather than ventricular activity that is targeted in RRI-based algorithms. It provides a patient specific detection of AF using a simple classifier, and can be leveraged as a tool to detect AF onsets/offsets over short AF episodes even when a patient's heart rate is controlled.
    Biomedical Signal Processing and Control 04/2015; 18. DOI:10.1016/j.bspc.2015.01.007 · 1.53 Impact Factor

Full-text (5 Sources)

Available from
Jun 4, 2014