How do I verify whether what I am removing from a noisy image is only the noise
1 Ansicht (letzte 30 Tage)
Ältere Kommentare anzeigen
farzad
am 8 Okt. 2018
Beantwortet: Walter Roberson
am 9 Okt. 2018
Hi All
I have to remove noise from an image, that I don't have a sample without noise to compare to. How could I know if I am not removing more than noise, like some data from the real image itself ?
0 Kommentare
Akzeptierte Antwort
Image Analyst
am 9 Okt. 2018
One way is to look at the difference of the denoised image and the original image.
diffImage = double(originalImage) - double(denoisedImage);
imshow(diffImage, []);
colorbar;
If you see any sort of structure in the image, then you're removing more than just noise. If it looks like completely random white noise with no spatial structures evident, then it's probably only noise that you've removed.
0 Kommentare
Weitere Antworten (1)
Walter Roberson
am 9 Okt. 2018
Create several images and corresponding noisy images with similar noise characteristics to those you are likely faced with, and adjust your code until you are satisfied that it only removes noise. Then apply it to your input image.
... Expect this process to take a number of centuries to come up with the right algorithm.
0 Kommentare
Siehe auch
Kategorien
Mehr zu Image Segmentation and Analysis finden Sie in Help Center und File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!