Working with Cell Arrays

I'm doing some image analysis and am having a problem working with the cell arrays. I read in my images and they are stored in a cell array. I want to threshold every image for the different channels. And later on work with those thresholded images.

However, with the code I have below, for channel1bw, channel2bw, etc. I end up with a 90112*2048 double instead of a cell array for each image in the channel. For this example there are 44 images for each channel 44*2048=9012. I'd like the thresholded images to be stored in a vertical cell array, instead of all the pixel data in one matrix.

Any help would be appreciated. Also is there a cleaner way to code this? The for loops are redundant and I know it is a poor way to code.

    channel1bw=[];
    channel2bw=[];
    channel3bw=[];
    channel4bw=[];
    channel5bw=[];
    for x=1:length(channel1)
        level=graythresh(channel1{x});
        channel1bw=[channel1bw; im2bw(channel1{x},level)];
    end
    x=0;
    for x=1:length(channel2)
        level=graythresh(channel2{x});
        channel2bw=[channel2bw; im2bw(channel2{x},level)];
    end
    x=0;
    for x=1:length(channel3)
        level=graythresh(channel3{x});
        channel3bw=[channel3bw; im2bw(channel3{x},level)];
    end
    x=0;
    for x=1:length(channel4)
        level=graythresh(channel4{x});
        channel4bw=[channel4bw; im2bw(channel4{x},level)];
    end
    x=0;
    for x=1:length(channel5)
        level=graythresh(channel5{x});
        channel5bw=[channel5bw; im2bw(channel5{x},level)];
    end
    x=0;

Thanks for any help.

[José-Luis]

What does

 size(channelx)

return?

[Jessica]

44x1 but depending on the image it could be nx1

[José-Luis]

I am sorry, but I don't follow. Does that mean that channelx is a cell array with images in it?

[Jessica]

yes, that is exactly what it is. Channelx is a cell array with the image's pixel data in it

Answer 1

[José-Luis]

your_mat = ones([2048,2048,numel(channel1)];
for x=1:length(channel1)
    level=graythresh(channel1{x});
   your_mat(:,:,x) = im2bw(channel1{x},level);
end

If you don't want to repeat the loops, you could make it a function.

Answer 2

[AJ von Alt]

You can use cellfun with an anonymous function to get rid of the for loop.

channel1 = { imread('cameraman.tif'),...
    rgb2gray(imread('ngc6543a.jpg')),...
    rgb2gray(imread('peppers.png')),...
    rgb2gray(imread('onion.png')) };

threshfun = @(x) im2bw( x , graythresh(x) );
channelbw = cellfun( threshfun , channel1 , 'UniformOutput', false );