how can i change the matrix dimension of an image?

when i am using edge detection for cameraman.tif(256*256)image the input will be of 256*256 uint8 the output is 256*256 uint8 but when lena.tif(256*256) the input size becomes 256*256*3 uint8 the output will be 256*768 uint8. for this reason i can evaluate psnr value for cameraman,but for lena image, the error is matrix dimension not match in between original image and edge detected image.can any one suggest for any code for solve this error?

Answer 1

[Walter Roberson]

Edge detection algorithms typically work on grayscale images. Some of them do work on color images, but the one you are working with is likely working on grayscale, so you will need to use rgb2gray() on the lena image before doing the edge detection.

[girish mishra]

Thank u sir.its working.can u give me the edge detection of canny operator and Log operator

Answer 2

[Image Analyst]

I don't know what your edge detection algorithm is doing but you might be using size() wrong. See Steve's blog: http://blogs.mathworks.com/steve/2011/03/22/too-much-information-about-the-size-function/

Also, lena is a color image and perhaps you need to convert it to grayscale with rgb2gray() or be extracting the green channel, which is usually the least noisy of the channels. Otherwise you'll have to decide how you want to compute and handle the psnr of a color image. By the way, the psnr() function in the Image Processing Toolbox can handle color images so is that what you used? (I doubt it.)

Answer 3

[John BG]

if you check size on the tiff file

A=imread('lena.tif')
[sz1 sz2 sz3]=size(A)

you will see the right size.

If you try

[sz1 sz2]=size(A)

MATLAB piles up everything that is not dimension 1 into the second size, giving a misleading size, that perhaps is what happens in your question

John

[e-mail]