A Novel Segmentation Methodology for Cursive Handwritten Documents

ABSTRACT

Segmentation in handwritten documents is a very challenging task because in handwritten documents curved text lines and non-uniform skews appear frequently. Basically segmentation means to analyse the document image into its sub component as text line, words, or ligatures and finally character. Various handwritten text recognition systems take segmented character as input to recognize them and some recognition systems which are based on holistic approaches use word level segmentation. In this paper, we have implemented a complete framework of segmentation for handwritten documents with various languages. A novel connectivity strength parameter is used for deciding the groups of the components which belong to the same line. Over-segmentation is also removed with the help of depth first search approach and iterative use of the connectivity strength function. We have implemented and tested this approach with English, Hindi, and Urdu text images taken from benchmark database and find that it is a language adaptive approach which provides encouraging results. The average accuracy of the proposed technique is .

Keywords:

• Connectivity strength function
• Cursive handwritten text
• Depth first search
• Segmentation

Additional information

Notes on contributors

Subhash Panwar

Subhash Panwar received his BTech degree in computer engineering from University of Rajasthan, Jaipur, India in 2004, MTech degree in computer engineering from Motilal Nehru National Institute of Technology, Allahabad, India, in 2010, and pursuing PhD degree from the Malaviya National Institute of Technology, Jaipur, India. He is working in the Government Engineering College Bikaner, Rajasthan, India as an assistant professor from September 2004. His current research interests include image fusion, computer vision, computational intelligence and pattern recognition.

E-mail: [email protected]

Neeta Nain

Neeta Nain is an assistant professor in the Department of Computer Science and Engineering, Malaviya National Institute of Technology, Jaipur. She has more than two decades of teaching experience. Her research area is image processing, pattern recognition and computer graphics. Presently she is guiding research in and written text recognition, corpus development, and crowd analysis. She has published a number of articles in international journals and conferences.

E-mail: [email protected]