Text line segmentation of offline malayalam handwritten document

Automatic handwriting recognition systems rely on text line segmentation. Text line segmentation faces a number of obstacles, such as irregular text line gaps, slanted text lines, cursive handwriting, and other similar problems. Furthermore, line segmentation is worse when dealing with noisy and unc...

Full description

Saved in:
Bibliographic Details
Main Authors: Anju, A. T., Chacko, Binu P., Basheer, K. P. Mohammad
Format: Conference Proceeding
Language:English
Subjects:
Documents
Handwriting recognition
Segmentation
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Automatic handwriting recognition systems rely on text line segmentation. Text line segmentation faces a number of obstacles, such as irregular text line gaps, slanted text lines, cursive handwriting, and other similar problems. Furthermore, line segmentation is worse when dealing with noisy and uncontrolled scanned handwritten manuscripts; dealing with Malayalam handwriting, which has a complicated style and character set, makes the process slightly more challenging. For the purpose of text line segmentation in unconstrained Malayalam handwritten scanned documents, this script independent technique primarily used morphological operations and contour extraction. More than that, the Hough line transform was used to fix the skewness of the scanned document picture. There were successful trials of the approach on ten other scripts, including Bengali, Devanagari, Gujarati, Gurumukhi, Kannada, Oriya, Roman, Tamil, Telugu, and Urdu.
ISSN:0094-243X
1551-7616
DOI:10.1063/5.0211999