Pre-segmentation in Offline Handwritten words
Main Article Content
Abstract
The paper deals with extraction of individual components from handwritten word image consisting of shadowed characters and defining criteria to identify touching and non-touching component in a word image. Touching characters need to be segmented appropriately in order to recognize. In case, if word contains touching characters, pre-segmentation and segmentation are the two required phases. Pre-segmentation assures that each segment should contain either a single character or touching components. The resultant segment need to identify as touched or isolated. Isolated component extracted from the word image can be recognized after feature extraction. Identified touching components of the image need further segmentation so that it can be segregated into individual components. Objects area criteria is used to embark upon the problem of shadowed characters. This paper also proposed analytic approach based technique- PCSW (Pixel Continuity Slope and Width technique) which help in differentiating between touching and non-touching(isolated) characters in a word image. The database for experimentation consists of legal amount words containing touching characters consisting of 1530 words by 15 different writers(db1) and 250 words dataset taken from database provided by ICDAR(db2). Implementation of PCSW achieved the accuracy of 98.16% and 96.80% on db1 and db2 respectively.
Article Details
Upon receipt of accepted manuscripts, authors will be invited to complete a copyright license to publish the paper. At least the corresponding author must send the copyright form signed for publication. It is a condition of publication that authors grant an exclusive licence to the the INFOCOMP Journal of Computer Science. This ensures that requests from third parties to reproduce articles are handled efficiently and consistently and will also allow the article to be as widely disseminated as possible. In assigning the copyright license, authors may use their own material in other publications and ensure that the INFOCOMP Journal of Computer Science is acknowledged as the original publication place.