how to separate out the lung sounds from heart sounds ?

i am doing my ME project in BMI please help me to sort out the prob.

Answer 1

[Wayne King]

Are these sounds separated in approximately disjoint frequency intervals? In other words, can you separate them in the frequency domain? In that case, you can use a filter.

[Chandrashekhar Choudhari]

Actually it is "Heart sounds localization using Lung sounds entropy" please help for the same

Answer 2

[Andreas Goser]

You need to provide FAR more information for a useful answer. "Separation" can't be the ultimate goal. Separation will be an intermediate step to do analasyis on lung or heart signal, right? It is also unclear if this is a post processing or real-time.

Approach: Assuming the heart signal has higher frequency than lung, not many changes in frequency (that depends on your application) and not much noise, I suggest to make a frequency analysis to identify heartbeat frequency and filter that out with a relatively low band filter.

[Chandrashekhar Choudhari]

Actually it is "Heart sounds localization using Lung sounds entropy" please help for the same

[Andreas Goser]

"You need to provide FAR more information for a useful answer."

[Chandrashekhar Choudhari]

Heart sounds are the main unavoidable interference in lung sound recording and analysis. With the signal processing and advances in computer technology, analysis of lung sounds becomes promising tool of pathology of lungs. Here the problem is to overcome with the heart sounds in lung sounds. Heartbeat is an unavoidable source of interference of lung sounds recording. When it occurs in lung sounds recording changes both frequency and time characteristic. Several methods based on adaptive filtering, wavelet denoising, adaptive thresholding and two-dimensional interpolation of LS in the time-frequency domain and removing HS-included segments from the wavelet coefficients of LS and then reconstructing the signal by auto regressive or moving average models have been proposed for HS reduction.

Answer 3

[Wayne King]

Hi Chandrasekhar, There are a number of free MATLAB toolboxes on the web for computing information-theoretic measures. The paper abstract you have copied and pasted above employs these measures. Perhaps you can obtain one of these free toolboxes and try to replicate the methods in the cited paper.

Here is one such toolbox:

http://www.cs.man.ac.uk/~pococka4/MIToolbox.html

Wayne

Answer 4

[Lucas García]

You can use adaptfilt in the DSP toolbox for adaptive noise canceling. A "similar" example can be found in the toolbox: separating maternal heartbeat from fetal heartbeat.