| Home | Organizers | Proceedings Editors | Proceedings Contributors | Search | | ||
|
|
|||
|
||||
|
|
||||
| Title: | DETECTION AND CORRECTION OF VERTICAL SKEW IN CHARACTERS | |
| DOI No: | 10.1142/9781860948534_0066 | |
| Source: | INNOVATIVE APPLICATIONS OF INFORMATION TECHNOLOGY FOR THE DEVELOPING WORLD (pp 423-434) | |
| Author(s): | T. VASUDEV
DoS in Computer Science, University of Mysore, Manasagangotri, Mysore, Karnataka, India G. HEMANTHKUMAR DoS in Computer Science, University of Mysore, Manasagangotri, Mysore, Karnataka, India P. NAGABHUSHAN DoS in Computer Science, University of Mysore, Manasagangotri, Mysore, Karnataka, India |
|
| Abstract: | In this paper, a method for detecting the vertical skew and correcting the skew for characters, specifically when subjected for recognition is presented. Character skew correction is an essential preprocessing event required during document understanding. If the skew in characters are not corrected then the readability of OCRs gets reduced. The method consists of three stages. In the first stage, the direction of skew is identified in character at a macro level with a knowledgebase. In the second stage the amount or angle of skew to its base line is detected with the support of line drawing algorithm. In the final stage, the detected skew in the character is corrected by transformation process. The method has been tested with various printed and handwritten samples and readability analysis is performed with respect to an OCR. | |
| Keywords: | character recognition; skew detection and correction; knowledgebase; line drawing algorithm; transformation process |
|
| Full Text: | View full text in PDF format (506KB) | |
| TOC: | Back to Table of Contents | |
|
||