Optical Character Recognition

1
Optical Character Recognition

• K.G.Srinivasagan M.E
• Asst. Professor,
• Dept. of CSE IT,
• National Engineering College,
• K.R.Nagar, Kovilpatti,
• Tamilnadu

National Engineering College
2

• Institution
• Inception 1984
• Approved by AICTE, New Delhi affiliated to Anna
  
• University, Chennai
• No. of UG courses 06
• No. of PG courses 07

National Engineering College
3

• Proposers
• Principal Investigator Dr.K.Ramar
• Prof. Head,
• Dept. of CSE IT
• Co-Investigator K.G.Srinivasagan
• Asst. Professor,
• Dept. of CSE IT
• Institution National Engineering College,
• K.R.Nagar, Kovilpatti,
• Tamilnadu
• Language Tamil

National Engineering College
4

• Name of the Component
• OCR package for Tamil handwritten documents
• Objective
• To develop a robust OCR for Tamil handwritten
  documents
• Scope
• Conversion of valuable handwritten Tamil
  documents of olden days into digital form as a
  part of promoting E-Governance.

National Engineering College
5

• Techniques Used
• Combination of Neural Networks Hidden Markov
  Model (HMM)
• Why?
• Neural Networks - powerful in discriminating
  shapes from different classes
• HMM - powerful in handling temporal
  sequence

National Engineering College
6
Proposed System
National Engineering College
7
National Engineering College
8

• Performance of the Techniques in other Languages
• English
• Tested found to be successful
• Estimate of Expected Performance
• 90 to 95

National Engineering College
9
PERT Chart
First Year Action Plan
National Engineering College
10
PERT Chart
National Engineering College
11
PERT Chart
National Engineering College
12

• Preferred Domains
• General Tamil handwritten documents
• Conventional hand written documents
• Predefined documents

National Engineering College
13
Thank You