Traditionally, the first line of a thesis is supposed to be some deeply meaningful and insightful statement meant to pull the reader in.
Oh well ...
This investigation analyses the data from an instrumented glove -- a
glove that returns information on finger position, hand position
and/or orientation -- for use in recognition of the signs that are
part of Australian Sign Language (Auslan). A system is developed for
recognising these signs, which is termed GRASP (Glove-based
Recognition of Auslan using Simple Processing)
.
The results will show that despite the noise and accuracy constraints
of the equipment used, reasonable accuracy rates were achieved. More
importantly, the techniques developed here can be applied to improved
hardware when (in accordance with an interface device version of
Moore's Law) it
becomes available. While GRASP may not be viable for everyday use, it
is possible that with improved instrumented gloves, future systems may
be.