American Sign Language Recognition Using Sensory Glove and Neural Network

Abstract

a system was designed to recognize
the Human hand manual alphabet of
the American Sign Language utilizing
artificial neural network that implement-
ed to convert the (ASL) finger spilling
alphabet into printed letters and using
matlab to implement program that con-
verting printed letters into animated
(ASL) the hardware system uses flex
sensors these sensors were positioned
on gloves to obtained The finger joint
angle data when represent each letter
of (ASL) and DAQ NI-6212 which was
the interface between the sensors and
the pc.DAQ produce 1000 readings
per second the average of this read-
ings is taken and normalization process
is performed on the data and then ap-
plied to trained neural network to rec-
ognize which letter was performed by
the hand and print it on the screen. On
the other hand a matlab program was
build using forward and inverse kine-
matics equations of human hand this
program take normal language letters
as input and produce an animated
(ASL) letter as output. The hardware
system have been trained and tested  for (ASL) manual alphabet words and names recognition.