Blind Source Separation from Multichannel Deconvolutive Mixtures

On this page are examples of separated murmurs and other artifacts from heart signals. Both the separation results and original signals can be listened to.

The techicalities of the separation approach are explained in Pietilä Astrid, El-Segaier Milad, Vigario Ricardo, Pesonen Erkki: Blind Source Separation of Cardiac Murmurs from Heart Recordings. ICA'06 Conference Proceedings.

Sporig data set

Listen to the above example of a speech artifact denoised from sporig data set.

Compare to the original data set: Channel 1 | Channel 2 | Channel 3 | Channel 4 | Channel 5 | Channel 6

14snd data set

Listen to the above example of a murmur denoised from 14snd data set.

Compare to the original data set: Channel 1 | Channel 2 | Channel 3 | Channel 4 | Channel 5 | Channel 6

20snd data set

Listen to the above example of an unknown artifact denoised from 20snd data set.

Compare to the original data set: Channel 1 | Channel 2 | Channel 3 | Channel 4 | Channel 5 | Channel 6

22snd data set

Listen to the above example of a murmur denoised from 22snd data set.

Compare to the original data set: Channel 1 | Channel 2 | Channel 3 | Channel 4 | Channel 5 | Channel 6