Automatic segmentation of content and noncontent based handwritten Marathi text document

Handwritten character recognition has been emerging topic studied in the last half century and shape up to the level which is sufficient to develop a technology driven application. Now the rapidly increase in the computation power, CR creates an increasing demand for new emerging applications, which require more advanced methodologies. The problem of character segmentation and its recognition in India is promising and challenging task because of variety of writing styles and complex structural features. The presence of the header line (shirorekha), overlapping of characters in the middle zone, touching characters, conjunct character makes the segmentation task more complex. The overall performance of an OCR system depends on segmentation process. Hence, in this paper an attempt is made to segmentation of content and Non content based handwritten Marathi text document.