segmenting feature fields

[David Harra]

I am trying to segment my data in such a way that the feature fields and and IPF are as close as they can be as I am trying to detect MTRs by propagating ultrasound through the structure.

As far as I can see, the two filters that affect the segmentation of the feature fields is :

1) Segment Features (C-axis Misalignment)

2) Minimum number of neighbours

I feel like I am just using trial and error and I am sure there is a logical explanation for the best method combination to use. Below is what my IPF and feature fields look like.

I was also wondering if someone can also explain what the intensity bar represents. The dark blue represents zero and this is data that is completely deleted as it doesn't satisfy the segmentation conditions  but what is the difference between the white colours and red?

In the images below I am using :

Minimum Number of neighbours =3

Segment Features (C-axis Misalignment) = 48

Thanks 

Dave :)