Differentiation of alphabets in handwritten texts

Our aim is to differentiate between parts of handwritten text written using different alphabets. We achieve our goal thanks to a fractal analysis of handwriting style. For each alphabet, a set of characteristics is extracted. Advantage is taken from the autosimilarity properties that are present in the handwriting. In order to do that, some invariant patterns characterizing the writing are statistically extracted. During the training step these invariant patterns appear while performing a fractal compression process, then they are organized in a reference base that can be associated with the alphabet. The alphabet identification is performed during a pattern matching process using the different reference bases successively. The results of this analyze are estimated through a correlation coefficient between the initial image of the text and a synthetic reconstruction of the text based on the references.