Writer identification from gray level distribution

When identifying a writer from a handwritten text, mostoften, either some characteristic patterns or some shapeparameters are extracted. They are assumed to be specificof the writer. Here, we are to explore a differentapproach, we consider the distribution of the pixel graylevels within the line. It is linked to pressure and writingspeed when text is realized.In the line, the direction that is perpendicular to thewriting way of drawing is privileged. The curve associatedwith the gray levels in a stroke section is characterized byuse of 4 shape parameters. More over the regular sectionsare selected and are grouped in section lots. Thedistributions of the sections and of the section lots arequantified. Thus 22 parameters are extracted. Threedifferent classifiers are used with and without geneticselection of the most significant parameters for theclassifier. Then the classifiers are combined and theresults show the gray level distribution within the writingis characterizing the writer in a significant way.