Investigating an Optimal Signal Epoch Length for Cardiotocographic Classification

Current Directions in Biomedical Engineering

Fuentealba Patricio¹  Illanes Alfredo²  Ortmeier Frank²  Poudel Prabal³ 

[1] Otto-von-Guericke University, po box 4120, 39106,Magdeburg, Germany;Otto-von-Guericke University,Magdeburg, Germany;Philips GmbH Innovative Technologies,Hamburg, Germany;

关键词: cardiotocograph;    foetal heart rate signal;    time-varying autoregressive spectrum;    iceemdan;   

DOI  :  10.1515/cdbme-2020-3132

来源: DOAJ

【 摘 要 】

This work focuses on investigating an optimal foetal heart rate (FHR) signal segment to be considered for automatic cardiotocographic (CTG) classification. The main idea is to evaluate a set of signal segments of different length and location based on their classification performance. For this purpose, we employ a feature extraction operation based on two signal processing techniques, such as the Improved Complete Ensemble Empirical Mode Decomposition with Adaptive Noise and time-varying autoregressive modelling. For each studied segment, the features are extracted and evaluated based on their performance in CTG classification. For the proposed evaluation, we make use of real CTG data extracted from the CTU-UHB database. Results show that the classification performance depends considerably on the selected FHR segment. Likewise, we have found that an optimal FHR segment for foetal welfare assessment during labour corresponds to a segment of 30 minutes long.

【 授权许可】

Unknown