How can i reconstruct a signal?

i have a signal (vector) with blanks (parts with zero value) and i want reconstruct the signal by fill the gaps with random parts of the main signals,how can i do this?

Answer 1

[Image Analyst]

Why not use interp1() or TriScatteredInterp(), or if it's an image, roifill()? Those all use neighboring values to fill in the blanks. Why would you want to use random parts of the signal to fill in the gaps? I don't see the sense in that.

[Image Analyst]

You can try this:

% Generate some sample data.
s = rand(1,30)
% Make blank between 10 and 20 and 26 and 29
s(10:20) = 0;
s(26:29) = 0

% Now do the task.
% Find the blank elements
blankElements = s == 0
numberOfBlanks = sum(blankElements)
nonBlankS = s(~blankElements) % just the non-zero elements
% Get random indexes.
elementsToTake = randperm(length(nonBlankS), numberOfBlanks)
% Assigm the elements of the non-blank s to the blank elements.
s(blankElements) = nonBlankS(elementsToTake)

[Samuel Tarazona]

Dear Sir, I would like to know how I can apply your script to the case in the figure above. The data I reconstructed is in the middle.

Thanks in advance!

[Image Analyst]

Looks like you already applied it. If you need more help, post your code and screenshot in a new question.

Answer 2

[Dimitris Malamas]

Thank you for the answer. I want to use random parts of the signal because it is a biosignal and this is the right way to fill the gaps.