Optical Handwritten Character Recognition for predesigned Forms

1
Optical Handwritten Character Recognition for
pre-designed Forms

• - Proposer -
• Prof. S. Rama Mohan
• Department of Applied Mathematics
• Faculty of Technology and Engineering
• The M. S. University of Baroda
• Vadodara 390002. Gujarat

Language Gujarati

• Components that will be implemented
• Template Generator
• Component Extractor
• Recognizer
• Verification with operator intervention
• Database / storage manager

2
Template Generator

• List of the technique (s) that will be used
• A software that will help operator to specify the
  details of the form
• XML document for recording the metadata
• Performance for these techniques in other
  Languages
• Successfully used by many professional packages
  like Fine Reader etc.
• Estimate of the expected performance (Quarterly
  PERT Chart)
• To be worked out
• Name of the domain for which the performance will
  be optimized
• Answer sheets of the objective question papers
• Other evaluation metrics in addition to the
  domain
• Forms for other type of complex data

3
Component Extractor

• List of the technique (s) that will be used
• Smearing and Connected component Extraction
• Performance for these techniques in other
  Languages
• Used with other kind of applications like OCR
• Estimate of the expected performance (Quarterly
  PERT Chart)
• To be worked out
• Name of the domain for which the performance will
  be optimized
• Forms with individual boxes
• Other evaluation metrics in addition to the
  domain
• Successful extraction while the character is
  broken or has two parts

4
Recognizer

• List of the technique (s) that will be used
• Performance for these techniques in other
  Languages
• Encouraging Results for using Fuzzy features in
  OCR Is available
• Template matching with Fringe Maps gives more
  than 90 character level recognition accuracy for
  Telugu script.(OCR)
• Estimate of the expected performance (Quarterly
  PERT Chart)
• To be worked out
• Name of the domain for which the performance will
  be optimized
• Neatly filled answer sheets with limited
  alternatives
• Other evaluation metrics in addition to the
  domain
• Reasonable recognition accuracy for answer sheets
  with eraser spots etc.

5
Verification with operator intervention

• List of the technique (s) that will be used
• A program displaying the image of the filled form
  and the recognized output side by side with the
  position where recognizer has shown poor
  confidence and would ask for operators input.
• Performance for these techniques in other
  Languages
• Used in many professional software like Fine
  reader.
• Estimate of the expected performance (Quarterly
  PERT Chart)
• To be worked out
• Name of the domain for which the performance will
  be optimized
• Forms with isolated boxes for each character with
  limited options
• Other evaluation metrics in addition to the
  domain
• Reasonable help for the forms without individual
  boxes for characters.

6
Database / storage Manager

• List of the technique (s) that will be used
• A module that will use JDBC / ODBC to store and
  retrieve the data to/from a database or file as
  per the user request.
• It will also help user to create and modify
  database
• Performance for these techniques in other
  Languages
• Used at various applications including this one
• Estimate of the expected performance (Quarterly
  PERT Chart)
• To be worked out
• Name of the domain for which the performance will
  be optimized
• Connectivity with MySQL database
• Other evaluation metrics in addition to the
  domain
• Connectivity with other databases like MS Access
  and with flat files.