Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Robust neural-network-based classification of premature ventricular contractions using wavelet transform and timing

Omer T Inan1, Laurent Giovangrandi, Gregory T A Kovacs

  • 1Department of Electrical Engineering, Stanford University, Stanford, CA 94305, USA. omeri@stanford.edu

IEEE Transactions on Bio-Medical Engineering
|December 13, 2006
PubMed
Summary

Related Concept Videos

Classification of Signals01:30

Classification of Signals

In signal processing, signals are classified based on various characteristics: continuous-time versus discrete-time, periodic versus aperiodic, analog versus digital, and causal versus noncausal. Each category highlights distinct properties crucial for understanding and manipulating signals.
A continuous-time signal holds a value at every instant in time, representing information seamlessly. In contrast, a discrete-time signal holds values only at specific moments, often denoted as x(n), where...

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

A pilot study of brain correlates of long-term treatment with transcutaneous vagal nerve stimulation in posttraumatic stress disorder.

Journal of affective disorders reports·2026
Same author

Quantifying changes in triaxial seismocardiography variability due to sub-optimal volume status.

Frontiers in physiology·2026
Same author

Exploring Variable Achilles Tendon Loading Via Active Acoustics.

IEEE sensors journal·2026
Same author

Establishing generalizability of wearable-enabled blood volume decompensation status estimation algorithms using transfer learning.

Computers in biology and medicine·2026
Same author

Non-invasive hemodynamic monitoring during hemorrhage and blood transfusion: opportunities and challenges.

Physiological measurement·2026
Same author

reSPIRE: A wearable multimodal sensing system for monitoring ventilation, cardiovascular dynamics, and respiratory muscle activity.

Biosensors & bioelectronics·2026

Accurate detection of premature ventricular contractions (PVCs) is crucial for diagnosing heart conditions. This study combined ECG morphology and timing data, achieving 95.16% accuracy in classifying beats, even on large datasets.

Area of Science:

  • Cardiology
  • Biomedical Engineering
  • Machine Learning in Healthcare

Background:

  • Accurate electrocardiogram (ECG) beat classification is vital for diagnosing heart conditions, particularly premature ventricular contractions (PVCs).
  • Existing algorithms often achieve high accuracy on small, overlapping datasets, limiting real-world applicability.
  • Significant variation in ECG morphology across patients reduces accuracy when using large datasets.

Purpose of the Study:

  • To develop a robust ECG classification method that maintains high accuracy on large, diverse datasets.
  • To improve the detection of premature ventricular contractions (PVCs) by integrating morphological and timing information.
  • To enhance the diagnosis of arrhythmias by improving beat classification accuracy.

Main Methods:

Related Experiment Videos

  • Combined wavelet-transformed ECG morphology with timing information to create a comprehensive feature set.
  • Utilized a neural network classifier trained on a subset of the MIT/BIH arrhythmia database.
  • Tested the classifier on both training and independent datasets comprising 93,281 beats from 40 files.

Main Results:

  • Achieved an overall classification accuracy of 95.16% across all 40 files (93,281 beats).
  • Demonstrated high accuracy of 96.82% on the independent test set (22 files not used for training).
  • Successfully differentiated between normal, PVC, and other types of heartbeats.

Conclusions:

  • Coupling ECG morphological information with timing data is effective for achieving high classification accuracy on large datasets.
  • The developed method shows significant promise for reliable, automated arrhythmia detection in clinical settings.
  • This approach addresses limitations of previous methods that struggled with inter-patient variability.