What kind of preprocessing I can apply for medical MRIs?

Hi,

I know my question may sound confusing, but I am facing with lack of the time.

I have MRI images of many patients that the intensity varies from patient to patient. I know that in some works they are applying histogram equalization and gamma correction, but guessing a correct value for the parameters of these functions is not straight forward. I did not have any choice except post my question here to get some expert advice.

My question is what kind of preprocessing I can apply on these MRIs that improves the processing and results? The following shows one slice: (The aim is to segment cartilage in MRIs)

or like this:

I am a beginner and do not know what kind of filters I can apply.

The other thing is that how can I find a proper parameter for functions. For example, I want to make the cartilage tissue to be appeared better than before, or I do not want to process muscles, so I want to their intensity to be decreased. How can I find optimal gamma value, for example, if I am using gamma correction?

I really appreciate your help.

Answer 1

[Asad Ali Siyal]

Hi Sara,

You can apply some basic thresholding or pre-processing techniques (such as brightness, contrast) for image enhancement. The further processing depends upon your interest, that what actually you are interested in?

Secondly, there are many filters (such as Gaussian, low-pass, high-pass, FFT) and function/algorithm for image segmentation or image registration that you can use effectively for performing image segmentation / extraction tasks.

[Sara Salimi]

Hi, Thanks for your comment.

I am actually confused what values I can set to these functions. for example brightness varies from patient to patient, How can I do histogram equalization for all patients with the same value.