Gaseous emboli detection based on a dual-wavelet transform analysis.

ABSTRACT

Emboli monitoring is nowadays based on the assessment of microembolic signals by Doppler ultrasound. However, the present systems have problems in detecting multiple emboli. A more dedicated algorithm for post-processing of the recorded Doppler signals was proposed. Based on the hypothesis that single and multiple gaseous emboli can be quantified by combining discrete and continuous wavelet transformation, the aim of this study was to detect gaseous emboli and to validate our method visually. A flow rig was used where gaseous emboli were generated. Doppler signals and visual validation data of gaseous emboli were acquired simultaneously. Microembolic signals were extracted and analysed using wavelet transformation. Results were validated against a visual reference. At various degrees of bubble generation, the system had 100 per cent detection during a low frequency of bubble generation but an estimation error of 7.4 per cent during a high frequency of bubble generation. The estimation error varied between -7.4 and +3 per cent. The system had a higher rate of success in detecting large gaseous emboli in small numbers than small gaseous emboli in large numbers. Single and double emboli were successfully detected and separated, whereas gaseous emboli clouds could be detected but not quantified. Being able to separate simultaneous gaseous emboli may offer new means of increasing detectability for embolism monitoring.