Multi Scale Image Processing Questions

Hi,

I'm new to multi scale image processing but I understand the general principle of doing it.

e.g. Making a pyramid of 2D images, enhancing each one and combining them back together.

Some online information suggested taking the original, applying a gaussian filter, then sub sampling it (quartering it), and doing this repeatedly until the image was 1 x 1 pixel. After this pyramid of images is created, each image in the pyramid is interpolated back to the original's size and subtracted from the original to produce a sub band image. Each sub band image is "combined " with the original to produce the final image.

I'm having some difficulty reproducing this with V13 (Using the API and OCX, with VB6). I can create the pyramid and subtract the images from the original and interpolate them back to the original's size. It's from here on I'm having problems. Maybe I don't understand the process as well as I thought.

I'm hoping some of you experts out there can provide a little assistance, particularly what should be done to each sub band (in general terms) and how to use LEADTools to properly combine them with the original.

Thanks in advance.

-PDW

If you want to use LEADTOOLS to implement multi-scale enhancement, the easiest way would be to use our MultiScaleEnhancement function. This function exists in LEADTOOLS Medical Imaging toolkits version 14 and later.

If you would like to try it before you consider upgrading, download the free evaluation edition of LEADTOOLS 16 from our site.