Rearrange matrix terms is a challenge

I am not finding exactly the right parameters to rearrange the terms of a matrix? It starts off as an RGB image I later convert to Lab, and want to store the Lab values to a text file in a particular rows x columns order, such that I end up with 49 rows by 33 columns. This is the code I have so far :

img = imread('Image 8-bit aRGBD65.tif');

imshow(img);

img_double = im2double(img);

img2Lab = rgb2lab(img_double, 'ColorSpace','adobe-rgb-1998', 'WhitePoint','d50');

LabIMG = reshape(permute(img2Lab,[3 1 2]),3,[]); % 3 x 1617

The LabIMG give mes a 3 rows x 1617 columns like this :

What I would like to have is for the data to be rearranged this way :

A1 = 42.29 66.60 0.69

B1 = some data (not shown)

C1 = some other data (not shown either)

... up to column 49 (AW1)

Then, the second row would follow with :

A2 = 47.57 58.15 1.54

B2 = ...

B3 = ...

.... Up to BW1

Third row :

C1 = 53.84 46.90 -0.31

C2 = ...

C3 = ...

... Up to CW1

Until I have 33 rows by 49 data points. I tried Transpose to no avail :

Transposed_CIELab = transpose(LabIMG);

I enclosed the resultant text file.... in case

I'm going in circles... Any help is appreciated.

[Stephen23]

@Roger Breton: please attach the image file.

[Roger Breton]

Here it is...

[Roger Breton]

I'm experimenting with a loop, but it's far from obvious...

[Stephen23]

imfinfo('Imge file.png')

ans = struct with fields:
                  Filename: '/users/mss.system.Rr7KOA/Imge file.png'
               FileModDate: '03-May-2022 20:20:05'
                  FileSize: 3546
                    Format: 'png'
             FormatVersion: []
                     Width: 49
                    Height: 33
                  BitDepth: 24
                 ColorType: 'truecolor'
           FormatSignature: [137 80 78 71 13 10 26 10]
                  Colormap: []
                 Histogram: []
             InterlaceType: 'none'
              Transparency: 'none'
    SimpleTransparencyData: []
           BackgroundColor: []
           RenderingIntent: []
            Chromaticities: []
                     Gamma: []
               XResolution: []
               YResolution: []
            ResolutionUnit: []
                   XOffset: []
                   YOffset: []
                OffsetUnit: []
           SignificantBits: []
              ImageModTime: '3 May 2022 20:20:05 +0000'
                     Title: []
                    Author: []
               Description: []
                 Copyright: []
              CreationTime: []
                  Software: []
                Disclaimer: []
                   Warning: []
                    Source: []
                   Comment: []
                 OtherText: []

Your image has 33x49x3 = 4851 elements. How do you expect to reshape 4851 elements into 33x49 = 1617 elements?

[Roger Breton]

I was coming to a solution on my own, using a loop. It's not elegant :

LabIMG = reshape(permute(img2Lab,[3 1 2]),3,[]); % 3 x 1617

% Note: Les données sont en ordre de Colonnes dans LabIMG

Position = 1;

RowStep = 1;

for i = 1:1617

Index(i) = Position;

Position = Position + 49;

reste = rem(i,33);

if (reste == 0)

RowStep = RowStep + 1;

Position = RowStep;

end

end

IndexColumn = transpose(Index);

CIE_L = transpose(LabIMG(1,:,:)); % --> k'th slice along x direction

CIE_a = transpose(LabIMG(2,:,:)); % --> k'th slice along y direction

CIE_b = transpose(LabIMG(3,:,:)); % --> k'th slice along z direction

Final = [IndexColumn, CIE_L, CIE_a, CIE_b];

Let me digest your comment...

At this stage, I'm going to read on Matlab "sorting". We'll see where it gets me but glad I thought about the remainder function, so that I can test the position in the row and determine whether I need to break it or not on the base of remainder. Sigh! ...

Answer 1

[Binaya]

Hi Roger, 

As per my understanding, you would like to merge the 3 channels of image in LAB format, given in text file into a single object to be put in an cell in the table.

Please find the below code for creating the table of size [33,49]:

load('Transposed CIE Lab.txt');
cellArray = cell(33, 49);
% Store each column in the cell array
for i = 1:33
    for j = 1:49
        cellArray{i,j} = Transposed_CIE_Lab((i-1)*10+j, :);
    end
end
table = cell2table(cellArray)

Following this you can export the created table into a text file.

I hope this solves your query.

Regards 

Binaya