A Novel Approach to Line Segmentation for Handwritten Gurumukhi Scripts
- Author Ravneet Kaur
- Co-Author Ashok Kumar Bathla
- DOI
- Country : India
- Subject : Engineering
Keywords : Mid-point, skewed, segmentation
Abstract
OCR (Optical Character Recognition) is used to read the digitally scanned documents, and convert the documents into computer readable or understandable format. While segmentation, (using OCR for digitally scanned documents) lines, words and characters are segmented. For computer written documents, lines are equally spaced, so it is very easy to segment the lines with higher rate of success. This success rate drops when the document is handwritten. Handwritten document may have lines not equally spaced. So there will be some difficulty in segmenting the lines from the document with higher efficiency. This research paper has proposed a technique based on mid-point. Where lines will be segmented based on mid-point. While segmenting the lines, skewness is also removed. This skewness can occur while scanning the document. Using mid-point approach the success rate of 94% has been achieved. This proposed technique has been applied on dataset of 500 images.
Comments
No have any comment !