ocr

23
OCR Algorithms Jacek Bajor

Upload: jacekb

Post on 04-Nov-2014

16 views

Category:

Education


2 download

DESCRIPTION

 

TRANSCRIPT

Page 1: OCR

OCR AlgorithmsJacek Bajor

Page 2: OCR

OCR

•Optical Character Recognition

•Translation of printed or handwritten text to digital form

KAZ 813

Page 3: OCR

OCR

•Fields

•Pattern Recognition

•Artificial Intelligence

•Machine Learning

•Computer Vision

Page 4: OCR

Applications

•Digitalising libraries

Page 5: OCR

Applications

•books.google.com

Page 6: OCR

Applications

•Office work

Page 7: OCR

Applications•Captchas

Mr. blocked

Page 8: OCR

Applications

•Recognising licence plates

•And many, many others...

Page 9: OCR

OCR Algorithms

1.Artificial Neural Networks

Page 10: OCR

Why ANN?

Page 11: OCR

Why ANN?

•They are flexible

•They can be thought

•They can learn themselves

•They are powerful

Page 12: OCR

ANNs

•Let’s assume ANN is a black magical box

Page 13: OCR

ANNs

•Learning stage

input outputaaaaaa

Page 14: OCR

ANNs

•Evaluation

98% a

5% a?

Page 15: OCR

ANN structure

Neuron (nerve cell)

Page 16: OCR

ANN structure

Page 17: OCR

ANN structure

Page 18: OCR

ANNs types

•Feedforward

•Recurrent (back-propagation)

•Radial basis function network

•Kohonen self-organizing network

•Others...

Page 19: OCR

ANNs applications

•Optical Character Recognition

•Function approximation

•Artificial Intelligence

•Finance

•Speech analysis

•Many, many others

Page 20: OCR

Advantages

•Flexible

•Efficient if well designed

Page 21: OCR

Disadvantages

•Focused on one activity

•Give approximations

•Inefficient if poorly designed

Page 22: OCR

Libraries

•FANN - Fast Artificial Neural Network Library - leenissen.dk/fann

•Annie - annie.sourceforge.net

•Libann - savannah.nongnu.org/projects/libann

Page 23: OCR

Thank you