Related Experiment Videos
Protocol to assess robustness of ST analysers: a case study
Franc Jager1, George B Moody, Roger G Mark
1Harvard-MIT Division of Health Sciences and Technology, 77 Massachusetts Avenue, Cambridge, MA 02139, USA. franc.jager@fri.uni-lj.si
Physiological Measurement
|July 16, 2004
Summary
This study introduces methods to evaluate ST segment analysis tool robustness against signal noise and database variations. Robust tools maintain performance across diverse conditions, ensuring reliable ST segment analysis.
Area of Science:
- Cardiology
- Biomedical Engineering
- Signal Processing
Background:
- ST segment analysis is crucial for diagnosing cardiac ischemia.
- Existing ST segment analysis tools may lack robustness to signal variations and database specificity.
- Standardized methods for assessing robustness are needed.
Purpose of the Study:
- To propose principles and methods for evaluating the robustness of ST segment analysers and algorithms.
- To define criteria for an ST analyser to be considered robust.
- To illustrate the application of the proposed protocol with a case study.
Main Methods:
- Developed an evaluation protocol including noise stress tests, bootstrap evaluations, and sensitivity analyses.
- Defined performance measures to assess robustness against input signal variation, database distribution, and parameter tuning.
- Applied the protocol to a Karhunen-Loève transform-based ST episode detection algorithm.
Main Results:
- The proposed protocol effectively assesses the robustness of ST analysers.
- The case study demonstrated the application of the robustness measures.
- Robustness is achieved when performance remains above critical boundaries during stress tests.
Conclusions:
- The presented principles and methods provide a framework for robust ST segment analysis.
- Robust ST analysers ensure reliable performance across different signal conditions and databases.
- This work contributes to the development of more dependable cardiac diagnostic tools.