Matlab multiple segments extraction

I am dealing with some matlab issue. Can someone help?

Suppose i have a sequence: A=[1 2 3 4 29 31 0 0 0 0 0 35 4 5 61 7 0 0 0 1 2 31 91 65], i need to extract nonzero segments seperately. For eg: my output should be s1=[1 2 3 4 29 31]; s2=[35 4 5 61 7]; s3=[1 2 31 91 65].

Answer 1

[Image Analyst]

If you have the Image Processing Toolbox, you can use PixelValues, which you can get from passing A into regionprops():

% Sample data:
A=[1 2 3 4 29 31 0 0 0 0 0 35 4 5 61 7 0 0 0 1 2 31 91 65];

% Make the measurements.  Get all the regions.
regions = regionprops(A>0, A, 'PixelValues')

% Pull them out from the structure into separate variables, if desired.
s1 = regions(1).PixelValues
s2 = regions(2).PixelValues
s3 = regions(3).PixelValues

In the command window:

regions = 
3x1 struct array with fields:
    PixelValues
s1 =
     1     2     3     4    29    31
s2 =
    35     4     5    61     7
s3 =
     1     2    31    91    65

[Sunil]

Great.. Thnx.. will install image processing TB...But what if i have this sequence??? i.e sequence which contains both positive and negative values? A=[1 -2 3 4 -29 31 0 0 0 0 0 35 -4 5 61 7 0 0 0 1 -2 -31 91 65];

[Image Analyst]

Use

regions = regionprops(A~=0, A, 'PixelValues')

[Sunil]

Thnx...

[Sunil]

@Image Analyst. Sorry my question was not clear earlier. My problem is: I have a sequence A=[1 2 3 -4 0 29 31 0 100 0 0 0 0 35 4 5 61 7 0 0 0 0 1 2 31 91 65]. Then i need to extract nonzero segments seperately like s1=[1 2 3 -4 0 29 31 0 100], s2=[35 4 5 61 7] and s3=[1 2 31 91 65] i.e A~=[0 0 0 0]. It should not break for one zero but it should break for 4 or more zeros.

[Image Analyst]

Use bwareaopen() to get rid of stretches of 3 or less.

% Sample data:
A=[1 2 3 -4 0 29 31 0 100 0 0 0 0 35 4 5 61 7 0 0 0 0 1 2 31 91 65]

% Find the zeros.
zeroLocations = A == 0
% Get rid of single zeros 0, pairs 0 0, or triples 0 0 0.
zeroLocations = bwareaopen(zeroLocations, 4)
% Find segments.  A segment may have single zeros.
nonZeroLocations = ~zeroLocations

% Make the measurements.  Get all the regions.
regions = regionprops(nonZeroLocations, A, 'PixelValues')

% Pull them out from the structure into separate variables, if desired.
s1 = regions(1).PixelValues
s2 = regions(2).PixelValues
s3 = regions(3).PixelValues

[Sunil]

Thank you.... it worked very well...

Answer 2

[Matt J]

Do you have the Image Processing Toolbox? If so, just use regionprops(...,'PixelIdx')

Otherwise,

http://www.mathworks.com/matlabcentral/fileexchange/24255-consecutive-vector-spliter

[Sunil]

It did not work as mine is not a repeated sequence... Its random.

Answer 3

[Sunil]

Is there a simple way to do it without using Image Processing Toolbox?

[Image Analyst]

Nothing simple comes to mind. You could try the brute force / intuitive way of using for loops to find the regions and using cell arrays to store the regions once you've found them. I guess you could consider that simple but it will be several lines of code.