binary image shown in MATLAB GUI Window (can not seen, all black)

Hi, there. I have a problem in showing the binary image in MATLAB GUI Window by using 'cData'. Codes are shown as follows:

    h_Img=handles.h_Img; %the image handle
    Img_BW=handles.Img_BW; % the binary image
    set(h_Img,'cData',Img_BW,'parent',handles.axes1);

This code lead to a black image. I cannot see the white part. But if I use this code

    Img_BW=handles.Img_BW;% the binary image
    figure;
    imshow(Img_BW,[]);

It will show good image. Anyone can help me? Thank you.

[Adam]

What is the data type of Img_BW?

[Penny]

Hi, Adam, data type of Img_BW is logical

Answer 1

[Image Analyst]

My guess is that Img_BW is actually uint8, not logical so imshow(Img_BW) would show it with gray levels 0 and 1, which essentially appears black, whereas doing imshow(Img_BW, []) will map the 1 to 255 so now it will appear.

The other option is that the white parts are so small that when your image is shrunk down to fit on the screen, they're essentially subsampled away. However in this case you would not see them with the [] option. Show me what this shows in the command window when you put these lines of code before imshow():

whos Img_BW
min(Img_BW(:));
max(Img_BW(:));

[Penny]

Hi,Image Analyst. It shows as follows:

    Img_BW      448x448            200704  logical
    logical 0
    logical 1

[Penny]

the image handle h_Img was got from code:

    h_Img=imshow(Img,[],'parent',handles.axes1);%Img is the original Dicom Image

In addition,the white parts are not too small. I can see the white parts in a smaller window. But in the GUI, I cann't see.

[Walter Roberson]

When you use imshow(Img) to create h_Img then if Img is much smaller than the Img_BW that you later set() as your CData, then the axes area will be too small and the new image will hardly appear. Using image() instead of imshow() is safer that way.

[Penny]

Hi, Walter Roberson. Img and Img_BW have the same size. They are the original Dicom Image and the corresponding binary image, respectively.

[Image Analyst]

If you're using set('cData') to bypass all the automatic adjustment stuff and put the image directly in there, and you had previously shown the 8 bit image in there, then putting an image with values of only 0 and 1 (your binary image) will probably still have the intensity range set up so that black = 0 and 255 = white (which is the default for 8 bit images). So putting 0 and 1 in there will show up as very dark/black. But when you use imshow() with [], it will adjust the intensity range (something that setting cData alone won't do) so then it will show up correctly.

Unless you're doing high speed imaging, like dozens of displays per second, I'd just use imshow() and not try to bypass the automatic stuff. I think setting cData alone is for like video processing, which I haven't heard you're doing.

[Penny]

All right. Thank you very much. Even though I am not working on the video processing, I do really have almost thousand images to be shown in the window. But at this moment, I will use imshow() again. Thanks so much.

[Image Analyst]

1000 is not a problem (anymore). With old versions of MATLAB, all those images would pile on top of one another and you'd have to call cla between calles to imshow(), but now it seems like only one image is there. If you need to do several per second, then you might use cData, but if it's just like displaying a new image every second or slower, then that's fine and you won't notice any speed difference.

[Penny]

Ok. Got it. Thank you very much.

Answer 2

[Walter Roberson]

colormap(gray(2))

[Penny]

Hi, thanks Walter Roberson. It doesn't work. I put the code like this:

    Colormap(h_Img,gray(2));
    set(h_Img,'cData',Img_BW,'parent',handles.axes1);

But it still shows nothing.

[Penny]

Oh,sorry, my updated code is:

    Colormap(handles.axes1,gray(2));
    set(h_Img,'cData',Img_BW,'parent',handles.axes1);

But it still shows nothing.