How to improve resolution of CT image?

Hi all

I want to improve the resolution of the CT images. Is there any filter that could improve the low-contrast resolution of medical CT images? I am working in Matlab. There are lots of filters for denoising. However, if the images are completely free of noise, the resolution also goes down. So I would like to go for filters that increases the resolution of the image. Your support will be much appreciated.

Thanks

Answer 1

[Walter Roberson]

No, you cannot improve the resolution of an image without extra information about what "should" be there. You are asking to do the equivalent of violating the Nyquist frequency.

You might be able to convert the information that is in the signal into a form more usable for further processing, but that would not involve increasing the resolution.

[Kavitha Srinivasan]

Hi

If that is the case, how can the CT images be used for diagnosing any soft tissue anomalies. I want to resolve objects that has similar density to that of background. My phantom has got a section for testing resolution. It has several circular discs of different diameters each has contrast difference of 0.01. we need to apply filter on that section and see if we could able to view them (currently we could not see those discs).

[Image Analyst]

A lot depends on the experience/skill of the technologist to know the proper parameters, such as kVp, etc. I'm not sure why you find it hard to believe that CT images that are routinely collected by the thousands every day would not be able to be used for diagnoses. It happens all the time and you know it does. Often though they will choose MRI over CT for really subtle soft tissue discrimination.

[Walter Roberson]

Perhaps you are asking about enhancing contrast differences rather than enhancing "resolution". "resolution", unless qualified, is a consequence of the signal sampling rate or physical parameters. I suppose one could loosely speak of increasing the visible "contrast resolution" within particular signal ranges.

[Aditi Deshpande]

Yes I think the OP meant 'contrast', not resolution. I also have this issue, I have CTA dicom images of the brain and in other DICOM viewers, the blood vessels are clearly visible but in MATLAB, upon using dicomread, the contrast is very poor and can barely see anything. Is there a solution for this?

[Bjorn Gustavsson]

When displaying image modify the gray-scale level with the caxis command?

If this is a case of nonuniform illumination effects one could try a homomorphic

transform - I've never gotten any improvements on my data with that one but

it is mentioned in introductory textbooks so perhaps someone can have a use for it.

[Walter Roberson]

Use the image contrast adjustment function to see where your interesting data is.