MATLAB Answers


How to Automatically remove noise using Image historgram based?

Asked by voxey on 23 Jul 2019
Latest activity Edited by Image Analyst
on 29 Jul 2019
How to Automatically remove noise using Image historgram based?


Can you elaborate the question with sample image?
Check this image. From this image, can we automatically remove noise?

Sign in to comment.

2 Answers

Answer by Image Analyst
on 23 Jul 2019
 Accepted Answer

I don't believe so, at least not via a global operation, since the noise is likely very close to the local means. So you can't just look at the histogram and say all bright stuff is noise. Perhaps you can use a scanning method where you're looking at the histogram just in a local area and looking for outliers, like a sigma filter or something. There are some very good noise filters out there. For example, you might want to check out imnlmfilt() to do a non-local means noise reduction.


Yes sir, but how do we distinct between noise and information in such images?
The filter makes that distinction, like the median filter assumes that the "signal" will be the median of values in the scanning window. If you have your own knowledge of what is signal and what is noise (even if it's probability distributions), then you can design a custom filter to take that into account.
I'm sure there are probably CT-specific papers on noise removal in medical images here in vision bibliography

Sign in to comment.

Answer by Ranjith veeran on 23 Jul 2019

  1 Comment

Ranjith, i am asking auto noise removal? your answer not related to my question ?

Sign in to comment.