Hi John,
The short answer is no.
3dSkullStrip started off per the description of BET algorithm in Steve Smith's paper. However, it has evolved a whole lot from that. For example, it uses 3dedge detection, eye avoidance, large ventricle avoidance, non linear expansion progression, a host of convergence criteria, and it looks inside and outside the surface as it is progressing. There is also the pre-processing that is done before 3dSkullStrip starts.
Those additions were made to make the program work well on 300 anatomical datasets of various MR signal quality, and anatomical conditions.
cheers,
ziad