I = im2double(imread('barbara.png'));
imshow(I);
s = size(I);
S = s(1) * s(2);
OriginalImageVector= reshape(I, S ,1);
p = 0.5; % p between 0 and 1
noisyimage = (I + p*rand(size(I)))/(1+p);
imshow(noisyimage, []);
s2 = size(noisyimage);
S1 = s2(1) * s2(2);
NoisyImageVector= reshape(noisyimage, S1 ,1);
When I compare the values of "OriginalImageVector" and values of "NoisyImageVector", the result is same why?
while when I added slat and pepper noise to the image, it shows some difference between the valuues of "SOriginalImageVector" and "SNoisyImageVector".
IS = imread('barbara.png');
noisyimageS = imnoise(IS, 'salt & pepper', 0.02);
figure; imshow(IS, []);
figure; imshow(noisyimageS, []);
s = size(IS);
S = s(1) * s(2);
SOriginalImageVector= reshape(IS, S ,1);
s2 = size(noisyimageS);
S1 = s2(1) * s2(2);
SNoisyImageVector= reshape(noisyimageS, S1 ,1);
Kindly explain when we added the Random Value Impulse noise to an image then how we can generate labels?
(in the slat & pepper noise case, I compared the noisy image values with the original image values and got the labels)
How do you compare the vectors?
Please be aware that gaussian white noise has expected value 0, therefore if you just sum errors the cumulated error could be very close to zero.
if You compute some absolute or squared errors then it would result in some notable difference.
Hi Shafaq, I run your codes, and it turned out the NoisyImageVector and the OriginalImageVector vectors are NOT the same. Have you actually compare the values of them? Use sum((sum(NoisyImageVector-OriginalImageVector))) to see if they're exactly the same (should not be zero).
Yes sir, I compared and both have "all" different values, means every pixel value is different of both vectors. From this way I only get the "Noisy labels. It means every pixel of noisy image is noisy? but it should not be as I am doing two class classification and in result I got only 1 class i.e. "Noisy"
There should be some noise free pixels of NoisyImageVector according to the literature.
- Badges
- Science topic
- Similar topics
- Computer Science
- Images