A discrete arabic script for better automatic document understanding

الملخص EN

This paper lays the groundwork for the development of new fonts to produce discrete Arabic script, for the first time, instead of cursive Arabic script.

These fonts help in automatic document understanding and can be used to print books, newspapers, periodicals, and all other printed materials.

Of course, all other properties of Arabic writing system are preserved when producing such fonts.

The history of Arabic calligraphy since its beginning provides a strong defense of our call to break the cursive law of Arabic script.

We could develop new fonts for discrete Arabic typography such that the characters can be segmented with simple vertical white cuts.

Two parameters are investigated to suit the new requirements: left and right white spaces.

Nine A4 pages of Arabic script were used in our experiments to empirically determine a sufficient amount of these spaces.

A font with left and right spaces of 160 FUnits each, achieved a segmentation success rate of 99.99%.