Computer displays nowadays support 16.7 million colors. It is 256*256*256, which means every color channel in a color image is represented in 8 bits. Matlab could do a double -> uint8 conversion automatically, but I'm sure, you wouldn't like the result. 

Welcome in the field of High Dynamic Range Imaging :) 

Computer displays nowadays support 16.7 million colors. It is 256*256*256, which means every color channel in a color image is represented in 8 bits. Matlab could do a double -> uint8 conversion automatically, but I'm sure, you wouldn't like the result. 

Welcome in the field of High Dynamic Range Imaging :) 

Thank you dear Gabor 

If i Understand it correctly,

Since double values requires 64 bits, total possible range of colors are (2^64)*(2^64)*(2^64). Computer Displays could not support such a range of color. 

For gray scale image it is not a problem as it normalizes the range.

use this:

imshow(uint8(image))

 Dear Hossein,

The question is about displaying an image without converting into unsigned integer. we can do it for   Gray scale image but can't for color image.

imshow('picturename.format') gives you the matrix form of the images only.

To Display the Double format Color Image (say A) using imshow use the following function

imshow(uint8(A))

S.Selva Nidhyananthan

I think the question is interpreted differently by many. I have attached the file with detailed explanation with an example, please read it.

The values of the color image have to be in the range 0-255 for uint8 images, and 0-1 for double.

Therefore, you have to transform the intensity values to the range 0-1 for imshow to work.

You can separate color image into three gray scale images as r , g, and b gray scale. Process then combine again in to color image then imshow it. 

As Andraz mentioned, for a double image to be color, it needs to be a 3 channel matrix with ranges between 0 and 1. By converting an image using double(im) you are keeping the range the same as before. In the case of a 3 channel uint8 image, this will result in most of the signals being larger than 1 and the results of imshow(double(im)) will be a predominantly (or completely) whitewashed image. To convert to double for display purposes, you need to divide by the maximum range of the original data type. For a uint8 image this would be imshow(double(im)/255). Best of luck.

your answer is simple technically when you transform image into double MATLAB did not  count it as image any more and you cannot use Imshow Function any more. because it is for images 

If the image was previously in uint8 and the values in range 0-255, then after conversion to double, you have to change the range to 0-1, by division of the image by 255. But you do not want to use im2double, thus I suggest that for some reason you want to keep the range 0-255. Thus you can simoly use 'imshow(A/255)' . Just remark: beware using some of the matlab image processing functions - most of them work in the way - if it is in uint8, then it expect range 0-255, if it is in double, then the expected range is 0-1. Therefore, you will getr a hell of trouble using the range 0-255 in double. Personaly, I am ALWAYS using conversion to double in the way 'A = double(A)/255' and ONLY when I need the  the range 0-255 in some function a call in in the way something = myfunction(A*255). Do not forget, that 'A = double(A)/255'  is not the same as 'A = double(A/255)' .

@Jan Urban

Thanks a lot....

/

Please try following command:

imshow(uint8(A));

where "A" is the input image which is already available with you in double format.

It's strange! When I do that ('imshow(A/255)' ) I get three back and white images. Can anyone attempt to explain why ?