An Arabic character recognition algorithm using 1-D slices of the char
acter spectrum is presented. The Fourier spectrum of the character's p
rojections on the X- and Y-axes is estimated. The features are extract
ed from this 2-D spectrum. The features of 10 sets of characters were
used as model features. The features of an input character are compare
d with the models' features using a distance measure. The model with t
he minimum distance is taken as the class representing the input chara
cter. Experimental results have shown that the presented algorithm is
capable of recognizing Arabic characters with a recognition rate of 99
.06%, using 10 features of the X-projection. This rate rises to 99.94%
when 10 features of the Y-projection are added. The proposed system w
as compared with another, based on the Fourier descriptors, which was
capable of recognizing 97.5% of test characters using 10 Fourier descr
iptors. The presented technique is superior to that of the Fourier des
criptors in terms of recognition rates and speed, as fast Fourier tran
sform is used in the calculation of the spectrum while standard equati
ons are used to compute the Fourier descriptors. Both techniques are i
nvariant to shift. However, the Fourier descriptor is invariant also t
o rotation and scale. (C) 1997 Elsevier Science B.V.