Under Construction

Alan Baydush's Research Page

I am currently a research associate in the department of Radiology at Duke University Medical Center. My research focuses on the use of statistical estimation techniques for the removal of scattered radiation in chest radiography and mammography. I will add links to some of my publications in the future as well as some sample images.!!

The current method that I use to remove scatter is by using a Bayesian image estimation technique to estimate the scatter free image based on a convolution model of scatter. Below, I show an image pair. The first image is an original portable patient image. The second image is a scatter reduced image. Both images look better than shown, but there was not much I could do about being limited to 8 bit greys per pixel. Our real images, actually, use 10 bits of information. Our measurements show that Bayesian image processing can reduce scatter without loss of resolution. This scatter reduction yields an increase in the contrast-to-noise ratio, which should result in an improvement in detection.

	This is a original measured patient image.
	This is a Bayesian proccessed image.

See the difference?? The processed image should appear sharper and clearer.

If you have comments/questions for me, please e-mail me by clicking here --> here