Online Handwriting Character Recognition For Sindh Language (MS Theses)
Laghari, Mehwish Supervisor Dr. Akhtar Hussain Jalbani
Online Handwriting Character Recognition For Sindh Language (MS Theses) - Nawabshah: QUEST, 2015. - 54p, :
ABSTRACT
The recognition of handwriting in general and recognition of cursive handwritten
script in particular grasped special attention by the research community working on
natural language processing. It presents more difficulties in the case of Sindhi
language (based on Arabic script), as it contains a richer character set of 52
alphabetic characters. In many cases the only difference between two characters is
the number, position or even the orientation of the dots, making it difficult to
identify the right character without context. In this research an application will be
developed for recognition of handwritten characters of Sindhi language. The
proposed application will work on characters entered on a touch sensitive device like
a PDA, smart phone or a tablet. In order to test the accuracy rate of the designed
application a richer dataset of 4000 samples has been collected by 60 writers. 20
writers were Primary School Teachers (PSTs ) and 40 other writers from various
fields of life. Then the proposed system is tested for accuracy results using that
dataset and various results have been presented and described in detail in this
research work.
xiii
Online Handwriting Character Recognition For Sindh Language (MS Theses) - Nawabshah: QUEST, 2015. - 54p, :
ABSTRACT
The recognition of handwriting in general and recognition of cursive handwritten
script in particular grasped special attention by the research community working on
natural language processing. It presents more difficulties in the case of Sindhi
language (based on Arabic script), as it contains a richer character set of 52
alphabetic characters. In many cases the only difference between two characters is
the number, position or even the orientation of the dots, making it difficult to
identify the right character without context. In this research an application will be
developed for recognition of handwritten characters of Sindhi language. The
proposed application will work on characters entered on a touch sensitive device like
a PDA, smart phone or a tablet. In order to test the accuracy rate of the designed
application a richer dataset of 4000 samples has been collected by 60 writers. 20
writers were Primary School Teachers (PSTs ) and 40 other writers from various
fields of life. Then the proposed system is tested for accuracy results using that
dataset and various results have been presented and described in detail in this
research work.
xiii