why I use imread fuction this photo not RGB array?

temp = imread('image.png');
% temp = 3000x4000 uint8?

why this photo don't have RGB 3D-array?

because I use pixelLabelDatastore function. then show result

Error using matlab.io.datastore.PixelLabelDatastore.label2cat

Pixel label image must have 3 channels when RGB-triplet pixel label IDs are specified.

Error in matlab.io.datastore.PixelLabelDatastore/label2categorical (line 1032)

L = matlab.io.datastore.PixelLabelDatastore.label2cat(...

Error in matlab.io.datastore.PixelLabelDatastore/readimage (line 578)

C = this.label2categorical(C, info);

It let me check the photo, why don't have rgb 3d array?

Answer 1

[DGM]

It's an indexed-color image. If you want it to be RGB, you can convert it.

% this is an indexed image and its corresponding color table
[inpict map] = imread('https://www.mathworks.com/matlabcentral/answers/uploaded_files/1485727/image.png');

% convert it to RGB
inpict = ind2rgb(inpict,map);

size(inpict)

ans = 1×3

3000 4000 3

imshow(inpict,'border','tight')

See also:

https://www.mathworks.com/help/matlab/creating_plots/image-types.html

[Latch Wu]

DGM, thanks your reply,

I has trouble in this code

    pxds = pixelLabelDatastore(labelDir,classes,labelIDs);
    C = readimage(pxds,1);

when I have many indexed-color image[s] in labelDir.

how to use pixelLabelDatastore function,

whether every indexed-color image file that must be transfer to RGB struction, then run function "pixelLabelDatastore" ?

[DGM]

I don't have CVT, so I'm not familiar with pixelLabelDatastore, but I think the way labelIDs are specified depends on the type of each image.

If it's an RGB image, they're specified as a color table, but if the file is indexed color, the labels are specified as the index.

% do it with an RGB input
classes = {'a','b','c','d','e','f'};
labelIDs = [0 0 0; 102 102 156; 128 64 128; 107 142 35; 0 0 142; 70 70 70];
pxds = pixelLabelDatastore('./',classes,labelIDs);
A = readimage(pxds,2);
size(A)

ans = 1×2

3000 4000

unique(A)

ans = 6×1 categorical array
     a 
     b 
     c 
     d 
     e 
     f 

% do it with an indexed input
classes = {'a','b','c','d','e','f'};
labelIDs = 0:5;

pxds = pixelLabelDatastore('./',classes,labelIDs);
A = readimage(pxds,1);
size(A)

ans = 1×2

3000 4000

unique(A)

ans = 6×1 categorical array
     a 
     b 
     c 
     d 
     e 
     f 

If the image type or classes vary within a directory, I see no way to handle that.