DETECTION OF EVENTS IN RESPIRATORY SOUNDS OF SUBJECT Russian patent published in 2024 - IPC A61B5/08 A61B7/00 

FIELD: medicine.

SUBSTANCE: group of inventions relates to medicine, namely to a method for identifying segments of a digital audio recording of respiratory sounds of a subject and a device for its implementation. Digital audio recording is filtered based on the characteristic frequency range of acoustic events to obtain a filtered digital audio signal. Filtered digital audio signal is processed to obtain a signal envelope. Samples making up the signal envelope are sorted into intervals to form a histogram. Histograms are approximated by statistical distribution. Threshold value for the signal envelope is determined based on the statistical distribution and a predetermined probability level. Signal envelope segments are identified, which exceed threshold value for identification of segments of digital audio recording as segments containing specific acoustic events. Determination of threshold value includes calculation of cumulative distribution function in relation to statistical distribution. Determination of threshold value includes determination of threshold interval, which corresponds to predetermined probability, level of which includes level of probability for cumulative distribution function.

EFFECT: reliable identification of events of interest in the analyzed signal, which contain counts with a value higher than background noise.

28 cl, 11 dwg