Efficient Binarization of Ancient Handwritten Vattezhuthu Documents

Vattezhuthu, mostly written on palm leaves, was one of the ancient scripts of India which existed from 8th century AD to 15th century AD. This was one of the degraded handwritten scripts due to many reasons. The main obstacles of binarizing these ancient handwritten texts are its variable background intensity, ink bleed, uneven illumination, blurred character and contrast issues, smear or smudge and shadows etc. Therefor an effective pre-processing method is required to binarize these degraded documents. In this paper, we propose a hybrid approach for binarization and enhancement of Vattezhuthu script. In the proposed technique pixel expansion is conducted on thin handwritten vattezhuthu script using super resolution. After that image enhancement is done using median filter. Noise is filtered out by finding out the contour of enhanced image by using contour approximation and area filter. Then the morphological operation, boundaries, is used on this result.