Detection and segmentation of text in handwritten Gurmukhi scripts

Abstract

Computerized character recognition has been an intensive and challenging research in the area of computer vision for many years. Such automated character recognition systems provide a solution for processing large volumes of data automatically. Character recognition could be broadly categorized into two sub-fields: Machine Printed Character Recognition and Handwritten Character Recognition (HCR). First, deals with the recognition of machine printed characters where second deals with the recognition of handwritten characters. HCR can be further divided into two sub-fields, namely, online HCR and offline HCR. Offline handwritten character recognition involves the development of computational method that can generate descriptions of the handwritten objects from scanned digital images. This is a challenging computational task, due to the vast impreciseness associated with the handwritten patterns of different individuals. To extract this imprecise knowledge, efficient algorithms of preprocessing, segmentation, classification are required for better recognition of characters, words and text lines. In the present work, we have focused on Punjabi handwritten text segmentation. Gurumukhi script is a two dimensional composition of symbols with connected and disconnected diacritics. Handwritten Gurumukhi script has some complexities like connected, overlapped text lines, words and characters. It is one of the major reasons for errors during the recognition process. In offline HCR, segmentation is a challenging job for writer independent handwritten document image. In this thesis, an algorithm is proposed for text line, word and character segmentation in handwritten Punjabi document. The proposed algorithm deals with the problems like overlapped and connected components in text lines and extracts text lines from handwritten document image. Later, words and characters are extracted.