With the advent of new fetal phonocardiographic transducers affording better signal qualities than hitherto possible, this non-invasive monitoring technique is explored with a view to long-term fetal heart rate (FHR) monitoring. The analysis of fetal heart phonocardiographic (phono) signals in spectral composition, stationarity and type of interference noise sets the basis for the development of FHR estimation algorithms. For repeatability purposes, the fetal heart signals are simulated at varying signal to noise ratios (SNR). This enables the controlled generation of simulated fetal phono time series which are then used to assess the performance of the investigated algorithms. The computation cost of each algorithm is extracted and given in processor independent format, thus giving an idea of the applicability to real-time FHR processing. A time domain beat-to-beat (BTB) FHR estimation algorithm, namely the 'Pseudo-Inverse Filter', is given, and its performance is subjectively compared to the pulse correlation approach. The pseudo-inverse filter is shown to offer increased accuracy in the detection of heart beat events. Contributions to block processing FHR estimation algorithms are also detailed in this work, notably the 'Comb Filter', 'Modified Block Autocorrelation' and the 'Iterative Block Clean-up'. The performances of fourteen FHR estimation algorithms are compared using varying SNR phono signals. Those algorithms based on block processing are further subjected to accelerating FHR signals to assess their tracking performances. Using pre-determined FHR inclusion bands, two measurable performance quantities are obtained: The outlier error rate (OER) and the in-band standard deviation (IBSD). Overnight recordings of phonocardiographic signals are processed in real-time using the 'Comb Filter' on a stand-alone, VME based, signal processing facility. FHR estimates obtained from two second blocks are then transmitted to the Mainframe computer for further processing and display.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:641364 |
| Date | January 1991 |
| Creators | Bassil, Hany Elias Milad |
| Publisher | University of Edinburgh |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | http://hdl.handle.net/1842/10745 |
Page generated in 0.0023 seconds