Non-invasive aneurysm detection using digital signal processing.
The attributes and performance expectations of a relatively low cost, non-invasive, aneurysm detection system have been studied. The developed system required that bio-acoustic data be captured at the patients' eyes, conditioned, transformed into the frequency domain, and analysed in some systematic fashion. It was admitted that the signal transducers were in need of improvement. Several low cost spectral analysers were built and tested. This system transformed the patient's bio-acoustic signal into a power spectrum. Features, representative of the patient's medical state, were extracted from the power spectrum and were input to a double-linked tree structure. Together with the medical class information, the structure can be searched to find the best statistical estimate of a patient's medical status. Experiment on generated and patient data, support the validity of the chosen diagnostic scheme.