Classification and digital restoration of damaged Arabic manuscripts

Abstract EN

This paper proposes a novel an effective approach to classify the damage in ancient Arabic manuscripts through damage manuscripts classification (DMC) model; into two types, each type has spatial model.

the first type is fading text color (FTC) model, and the second is missing part of text (MPT) model.

for the first type, which is the fading text color, it is done through (FTC) model, where segmentation, contour strength and contour size algorithms were applied.

as for the second type, we applied some segmentation algorithms to separate image of damaged manuscripts into foreground and background.

segmentation by thresholding applied on background to detect if there are missing in text to complete it based on database which we had prepared for this purpose including the handwritten arabic fonts and the forms of letters in different styles such as Naskh, Reqaa,..etc.

detection of style algorithm is also used to determine the style of missing text according to the same database, then digital restoration applied to the image.

applying pre-processes on the used data yields good classifications’ results.

the contribution of this work is the introduction of synthetic features that enhance the classification performance.

for testing purposes, two famous books from the Islamic literature are used: 1) pages of the ottoman Quran and 2) some Quran verses in Naskh script.