Dorner and Hagen are unique in that they have taken a holistic approach to the question of ASL interpretation [DH94]. Dorner's Master Thesis built a glove that would capture information about hand movement, while Hagen's Master Thesis involved building a deductive database that successfully translates from a standardised form of ASL into spoken English.
Dorner uses a cotton glove, with various areas of it painted different
colours to enable tracking. Unlike most other models used by other
people in this field, Dorner attempted to construct a
3+1D
model of the hand. Each joint is marked by three
rings: A central band, indicating the finger it belongs to, and two
bands of the same colour -- either side of the central band -- indicating
which joint it is. To improve performance, a predictive model was
emploted, limiting the possibilities of movement to those that are
humanly possible. This approach appeared to work reasonably well;
however the calculations involved became complex, stopping it for
being real time in practical cases.
Hagen in the meantime, created a deductive database for ASL, adding the necessary features for purposes of indication which are above the lexical level -- things like raising the eyebrows to indicate a question, and even allow 5 spaces for reference to objects (a technique often used by signers).
They then began to work on putting these two systems together, as described in [DH94]
Although not complete yet, they hope to connect these two systems together to form a complete ASL system. There is of course a great deal of filling to be put between these two slices of work.