I'm using hexagon.

here's what I was looking at in the OBJ files...
File Au  (unmorphed, UV Mapped)
begining of the file:

#Wavefront OBJ file created by Hexagon 2
mtllib apron-zero-uv.mtl
g Apron

4572

v -0.06199 0.43279 -0.01487
v -0.06185 0.43237 -0.01504
             ...etc
             ...then switch from v to vt.. 
v -0.04615 0.58628 0.04128

4709 1

vt 0.61878 0.922992
vt 0.618964 0.936212
             ...skipping normals...then switch from vn to f..
vn -0.167595 0.598853 -0.783127
vn -0.167595 0.598853 -0.783127
f 323/325/352 324/326/353 303/314/330 302/310/329
f 323/325/352 302/310/329 300/306/327 321/324/350

I see the in my grouped file, the grouped vertices start with a particular coordinate v, and continue in the same order as they did in the other file. Let's say group 'neck' started at vertex 62 of the original file Au, and continued for 125 vertices (obj file text lines)... in this file.  I'm seeing that in my files Bu, Cu.. the same thing happens. The neck vertices always start at vertex 62, and go for 125 lines.  Same for other groups, 126+62nd line always gives me a vertex that starts a chest group. In each of the files.

This leads me to believe that one can write a utility that does the following:
Compare grouped and ungrouped object Au and Ag, and determine that...
Vertex 1-325=neck group *(I'm making up the numbers here, for illustration, the program would determine the real numbers by searching for mesh coordinate identical to the begining of each group)*Vertex 326-1487=chest group
Vertex 1488-2579=skirt group
Vertex 2580-3711=hip group etc...

now, take one of the morphed objects (with identical vertex order) and chop up the text.
Take verteces 1-325, and re-write them in a new obj file with it's relative group heading,
take the next set of vertices, and write out the next group..... 

And one ends up with a regrouped obj file, based on a grouped template.

I'm talking about vertices only here. To create morphing delta's for inj, or for inclusion in cr2, I only need xyz information for vertices. 
I bet same, or similar thing could be done for facets too, as I understand that facets depend on a particular order too, but I didn't analyze it on an actual sample yet. 

If you're interested in getting the copy of the obj files I'm talking about, please email me at conniekat8@yahoo.com, and I'll be happy to send them.

Hi, my namez: "NO, Bad Kitteh, NO!"  Whaz yurs?
BadKittehCo Store  BadKittehCo Freebies and product support