Ali Hadi Hussein1, Ashor Al Sellami2
1High Vocational Comperhensive Institute Bengashier Tripoli – Libya
2Faculty of Engineering -Al Fateh University -Tripoli – Libya
In this work, a system that has the ability to recognize and identify persons from their
handwriting is designed. Wavelet transform algorithm was utilized to analyze the individual
characters of certain person's writing. Complex moments are used as tool for the
differentiation between different persons.
This involves process that requires isolating the individual characters from the input images,
because adopted writing involved series of characters; e.g. capital English letters "A through
The 1st order decomposition level of the wavelet transform is utilized. This utilization is
adopted to perform the recognition and identification process quickly and adequately.
The test was implemented on 40 samples of 10 persons voluntaries. Some of the samples were
used for verification tests.
Minimum distance classifier (based on Euclidian distance) is used to test the similarity or
closeness between matched samples. Three handwriting versions, for each person, has been
adopted and used as reference data reserved in the constructed database.
Complex moments of four decomposed sub images (i.e. LL, LH, HL, HH) were computed
and used as featured vectors to be compared in the recognition and identification processes.
The developed system utilized the power of the Visual Basic 6 for performing the required
operations involved in our designed system.