Abstract
Ventricular arrhythmias (VA) are dangerous pathophysiological conditions affecting the heart which evolve over time resulting in different manifestations such as ventricular tachycardia (VT), organized VF (OVF), and disorganized VF (DVF). Success of resuscitation for patients is greatly impacted by the type of VA and swift administration of appropriate therapy options. This thesis attempts to arrive at computationally efficient, data driven approaches for classifying and tracking VAs over time for two purposes: (1) ‘in-hospital’ scenarios for planning long-term therapy options, and (2) ‘out-of-hospital’ scenarios for tracking progression/segregation of VAs in near real-time. Using a database of 61 60-s ECG VA segments, maximum classification accuracies of 96.7% (AUC=0.993) and 87% (AUC=0.968) were achieved for VT vs. VF and OVF vs. DVF classification for ‘in-hospital’/offline analysis. Two near real-time approaches were also developed for ‘out-of-hospital’ VA incidents with results demonstrating the high potential to track VA progression and segregation over time.
Highlights
THIS chapter will Ventricular arrhythmias (VA) previously explore the to provide various context signal analysis for this work.techniques that have been used to study It will introduce and explain the specific techniques that will be used in this thesis
That made those features suitable to classifying organized ventricular fibrillation (VF) (OVF) and disorganized VF (DVF), but less ideal for ventricular tachycardia (VT) and VF. It is well-known that VT and VF may be accurately separated by frequency parameters, and in this chapter the instantaneous mean frequency (IF) and IB2 features helped in that classification
CARDIOVASCULAR conditions known as VAs may seriously impact the health of patients and could even become lethal
Summary
THIS chapter will VAs previously explore the to provide various context signal analysis for this work.techniques that have been used to study It will introduce and explain the specific techniques that will be used in this thesis. THIS chapter will VAs previously explore the to provide various context signal analysis for this work. The first approach uses quadratic time-frequency energy distributions (TFDs), in which a signal is transformed into an energy map that is a function of time and frequency. These distributions typically have high TF resolution, useful for visualization and extraction of complex instantaneous features, but have high computational complexity [12], [22]. These methods contain cross-term artifacts in the TF plane, disturbing the interpretation of the signal at hand. A signal is approximated by a number of TF basis functions that are translated and scaled to match the structure of the signal
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
