Wavelet Based Denoising and Cross Correlation for Identifying Features in CTG Traces
Dovydas Kulcickis-Gutas, William Peter Marnane · 2022
This work aims to demonstrate a method of pattern identification in CTG traces. Following de-noising using the discrete wavelet transformation, the zero normalized cross-correlation is applied between the trace and the pattern of interest. This study allows a custom pattern to be cross correlated with a CTG trace, and areas of maximal overlap are indicators of successful identification. The method has been tested with CTG records from Cork University Maternity Hospital to determine the prediction rate.