Time-Domain Fragment-Similarity Processing of the Auscultation Records for Detection of their Periodic Components

摘要

The ability to identify the inherent rhythmic pattern of life activity sounds recorded by means of electronic stethoscopes or phonospirographic complexes is important for an efficient analysis of the auscultation data. For determining of properties and types of modulation periodicity present in such signal, a technique based on finding of similarity of its fragments is proposed. In doing so, the reference fragments of the record are shifted along time axis, compared with the current ones and the difference in their acoustic power is estimated. The presented examples show that such processing of respiratory and cardiac sounds can provide the successful simultaneous tracking of both the rhythmics of the heartbeat and respiratory cycle. Moreover, the irregular zones generated by the fragments that do not have got similar ones over the entire record are identified in the resulting graphic portraits of acoustic signals.