Delete items from array and split it up into new arrays

Hi,

I have a long array consisting of a long section of negative values, then a long section of positive values, then a long section of negative values and so on. I would like to remove all the negative values from my array, but at the same time split up the array into the individual long sections of positive values.

Simplified example:

array = [-0.0016 -0.0017 0.1746 0.1834 -0.0017 -0.0018 0.1567 0.1456 0.1744 -0.0018 -0.0017 0.1243 0.1456]

And then I would like this:

answer = {0.1746 0.1834} {0.1567 0.1456 0.1744} {0.1243 0.1456}

So the negative values has been removed, and the array is split up into new arrays containing the positive values from each section.

I hope it makes sense!

Answer 1

[Image Analyst]

You can easily do it if you have the Image Processing Toolbox.

array = [-0.0016 -0.0017 0.1746 0.1834 -0.0017 -0.0018 0.1567 0.1456 0.1744 -0.0018 -0.0017 0.1243 0.1456];
% Identify positive indexes
mask = array >= 0;
% Measure values at those indexes.
positiveRegions = regionprops(mask, array, 'PixelValues');
% Identify negative indexes
mask = array < 0;
% Measure values at those indexes.
negativeRegions = regionprops(mask, array, 'PixelValues');

positiveRegions and negativeRegions are structure arrays (rather than cell arrays). To extract the numbers for a particular region, say the 2nd region you can do

values = positiveRegions(2).PixelValues

values = 1×3

0.1567 0.1456 0.1744

You can also get them in a table if you want:

array = [-0.0016 -0.0017 0.1746 0.1834 -0.0017 -0.0018 0.1567 0.1456 0.1744 -0.0018 -0.0017 0.1243 0.1456];
% Identify positive indexes
mask = array >= 0;
% Measure values at those indexes.
positiveRegions = regionprops('table', mask, array, 'PixelValues')

positiveRegions = 3×1 table
          PixelValues       
    ________________________

    {[       0.1746 0.1834]}
    {[0.1567 0.1456 0.1744]}
    {[       0.1243 0.1456]}

% Identify negative indexes
mask = array < 0;
% Measure values at those indexes.
negativeRegions = regionprops('table', mask, array, 'PixelValues')

negativeRegions = 3×1 table
       PixelValues    
    __________________

    -0.0016    -0.0017
    -0.0017    -0.0018
    -0.0018    -0.0017

[Bjælle]

That's perfect, thank you!

Answer 2

[Voss]

array = [-0.0016 -0.0017 0.1746 0.1834 -0.0017 -0.0018 0.1567 0.1456 0.1744 -0.0018 -0.0017 0.1243 0.1456];

is_negative = array < 0

is_negative = 1×13 logical array
   1   1   0   0   1   1   0   0   0   1   1   0   0

start_idx =  strfind([true is_negative],[true false])

start_idx = 1×3

3 7 12

end_idx = strfind([is_negative true],[false true])

end_idx = 1×3

4 9 13

answer = arrayfun(@(s,e)array(s:e),start_idx,end_idx,'UniformOutput',false)

answer = 1×3 cell array
    {[0.1746 0.1834]}    {[0.1567 0.1456 0.1744]}    {[0.1243 0.1456]}

[Bjælle]

Thank you so much!

[Voss]

You're welcome!

Answer 3

[AH]

You may find the code below doing the desired task

array = [-0.0016 -0.0017 0.1746 0.1834 -0.0017 -0.0018 0.1567 0.1456 0.1744 -0.0018 -0.0017 0.1243 0.1456]'

array = 13×1

-0.0016 -0.0017 0.1746 0.1834 -0.0017 -0.0018 0.1567 0.1456 0.1744 -0.0018

% find the mask where the array is non-negative
binMask = array >= 0;

% Find the start and end index of each region of interest
sidx = find(diff(binMask) == 1)+1

sidx = 3×1

3 7 12

eidx = find(diff(binMask) == -1)

eidx = 2×1

4 9

if (binMask(1) == 1)
    sidx = [1;sidx];
end
if (binMask(end) == 1)    
    eidx = [eidx; length(array)];
end

% Split it into cells
ncell = min(length(sidx),length(eidx));
A = cell(ncell,1);
for i = 1:ncell
    A{i} = array(sidx(i):eidx(i));
end
A

A = 3×1 cell array
    {2×1 double}
    {3×1 double}
    {2×1 double}

You can also achieve this task using the functions sigrangebinmask and binmask2sigroi.

[AH]

Here is the code using sigrangebinmask and binmask2sigroi

% Find mask 
threshold = 0; % split array into negative and non-negative parts
l = sigrangebinmask(array,threshold)
roilims = binmask2sigroi(l)

sidx = roilims(:,1) % beginning index
eidx = roilims(:,2) % ending index

ncel = size(roilims,1)

[Bjælle]

Thank you very much!

[Image Analyst]

@AH, what toolbox is sigrangebinmask in? It's not in the Signal Processing Toolbox.