Q1: The image is broken down into 2D frequency components in a pyramid fashion through Gaussian and LaPlacian tranforms and filters, and then each frequency component can be manipulated separately before being recombined to create the "improved" full image. Is this what you are doing?
A1: Yes.
Q2:You also have a parameter which can be set to Gaussian, bi-cubic, re-sample or normal for the filter. Are these filters applied at each scale?
A2: Yes.
Q3: How do you determine how to process different scales differently.
A3: All the scales are treated by the same manner.
Q4: what does changing the level of edge or level of latitude do. Can you send me more detailed information on this method and the parameters associated with it?
A4: Edge Level in general enchances the edges of the big features like bone, skull...etc, the more levels you set, the more frequency components this enhancment will be applied to.
Latitude Level is general enchances the soft feature of the image like tissues...etc, the more levels you set, the more frequency components this enhancment will be applied to.
For more information about Edge/Latitude Levels, please refer to the documentation.