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
  1. Home
  3. A Comprehensive Inference-time Augmentation Framework In Physiological Signals: Application To Ppg-based Af Detection.
  1. Home
  3. A Comprehensive Inference-time Augmentation Framework In Physiological Signals: Application To Ppg-based Af Detection.

Related Concept Videos

You might also read

Related Articles

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

Sort by
Same author

A Comprehensive Inference-Time Augmentation Framework in Physiological Signals: Application to PPG-Based AF Detection.

Physiological measurement·2026
Same author

Predicting Post-Radiotherapy Epigenetic Age Acceleration From Pre-Treatment Data Using a Machine Learning Framework in Head and Neck Cancer Patients.

Cancer medicine·2026
Same author

Bilateral Cingulotomy as a Therapeutic Option for Intractable Cancer Pain: a Systematic Review.

Current pain and headache reports·2025
Same author

Efficacy and Emerging Role of Erector Spinae Plane Block for Pain Management during Labor.

Current pain and headache reports·2025
Same author

Continuous Cardiac Arrest Prediction in ICU using PPG Foundation Model.

Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Annual International Conference·2025
Same author

DietAI24 as a framework for comprehensive nutrition estimation using multimodal large language models.

Communications medicine·2025
Same journal

When is Enough Enough? A Proposed Termination Point for the Number of Replicates in Computational Simulations.

ArXiv·2026
Same journal

Spatially Masked Regression Reveals Local and Distributed Predictability in Electrophysiological Recordings.

ArXiv·2026
Same journal

A beam--membrane biomechanical vocal fold model incorporating posturing and glottal conformation.

ArXiv·2026
Same journal

Analyzer-less X-ray Interferometry with Super-Resolution Methods.

ArXiv·2026
Same journal

Maximum Matching Accuracy: An Instance Segmentation Evaluation Metric Utilizing Globally Optimal Matching.

ArXiv·2026
Same journal

Stresses and fluid flow in lamina cribrosa through anisotropic poroelasticty.

ArXiv·2026
See all related articles

Related Experiment Video

BrainBeats as an Open-Source EEGLAB Plugin to Jointly Analyze EEG and Cardiovascular Signals
08:22

BrainBeats as an Open-Source EEGLAB Plugin to Jointly Analyze EEG and Cardiovascular Signals

Published on: April 26, 2024

A Comprehensive Inference-Time Augmentation Framework in Physiological Signals: Application to PPG-Based AF

Davood Fattahi, Runze Yan, Saurabh Kataria

    Arxiv
    |June 24, 2026

    View abstract on PubMed

    Summary
    This summary is machine-generated.

    Inference-time augmentation (ITA) enhances physiological signal classification by applying optimized transformations during real-time analysis. This model-agnostic approach improves accuracy for conditions like atrial fibrillation detection, even without retraining.

    More Related Videos

    Concurrent EEG and Functional MRI Recording and Integration Analysis for Dynamic Cortical Activity Imaging
    11:28

    Concurrent EEG and Functional MRI Recording and Integration Analysis for Dynamic Cortical Activity Imaging

    Published on: June 30, 2018

    Related Experiment Videos

    BrainBeats as an Open-Source EEGLAB Plugin to Jointly Analyze EEG and Cardiovascular Signals
    08:22

    BrainBeats as an Open-Source EEGLAB Plugin to Jointly Analyze EEG and Cardiovascular Signals

    Published on: April 26, 2024

    Concurrent EEG and Functional MRI Recording and Integration Analysis for Dynamic Cortical Activity Imaging
    11:28

    Concurrent EEG and Functional MRI Recording and Integration Analysis for Dynamic Cortical Activity Imaging

    Published on: June 30, 2018

    Area of Science:

    • Biomedical Engineering
    • Signal Processing
    • Machine Learning

    Background:

    • Physiological signal classification faces challenges from noise, artifacts, and data shifts in real-world applications.
    • Inference-time augmentation (ITA) offers a model-agnostic solution to improve robustness without retraining.
    • Current ITA methods for physiological signals are limited in scope and parameter optimization.

    Purpose of the Study:

    • To introduce a unified framework for inference-time augmentation (ITA) tailored for physiological signals.
    • To address the limitations of existing ITA methods by incorporating diverse transformations and optimized parameters.
    • To enhance the reliability of physiological signal classification in deployment settings.

    Main Methods:

    • Developed a framework integrating 13 augmentation techniques across time, amplitude, and frequency domains, plus artifact injection.
    • Utilized Bayesian optimization to fine-tune hyperparameters for these augmentation methods.
    • Evaluated the framework on atrial fibrillation (AF) detection using photoplethysmography (PPG) signals with GPT-PPG and ResNet models across multiple datasets.

    Main Results:

    • Standard ITA consistently improved Area Under the Receiver Operating Characteristic Curve (AUROC) and Area Under the Precision-Recall Curve (AUPRC) for AF detection.
    • GPT-PPG and ResNet models showed significant improvements in AUROC (up to 8.5% and 0.7%) and AUPRC (up to 10.6% and 0.8%).
    • Selective ITA further reduced the False Positive Rate (FPR) by up to 4.4% on non-AF datasets.

    Conclusions:

    • Inference-time augmentation (ITA) is a practical, model-agnostic strategy for enhancing PPG-based AF classification reliability.
    • The proposed framework effectively improves classification performance in deployment scenarios where retraining is impractical.
    • ITA demonstrates broad applicability for improving the robustness of various physiological signal analyses.