I've actually just done the same thing myself.
I'm not sure if BTK can get the segment angels if your motion capture model doesn't already calculate it (maybe someone else can answer this??)
However,if you can extract relative joint angles, that means your model most probably outputs joint center locations w.r.t global /lab coordinate system. From there, segment angles can be calculated from taking inverse tangent of differences between two adjacent joint centres, I.e. If knee joint center is a=[x1,y1,z1] and ankle joint is b=[x2, y2, z2],both w.r.t lab coordinate system, then shank segment angles in one of the directions is theta = atan[(x1-x2)/(z1-z2)].
This is just one approach I used - not validate at all. So I'd love to hear others' approachs or maybe btk has something built in already!
Kind regards,
Victoria