Umu logo Umeå universitet
Teknisk-naturvetenskaplig fakultet
Institutionen för datavetenskap


Assignment 3

Due date for the assignment is march 18, 2005

Background

Information input to modern computer artifacts more and more leave the traditional keyboard for the benefit of easier, smaller or faster input solutions. Some concrete examples of such artifact are small hand held computers, where Palm and the character recognition pen, C-Pen are the most popular ones. These artifacts involve classify handwritten and printed characters, such as letters and numbers. The recognition (extraction and classifying) of printed characters is called OCR (optical character recognition). OCR-software is also often delivered together with a scanner used to convert printed document to computer files.

Assignment

Your assignment is to implement an automatic OCR-solution. This means that you should extract written numbers and classify these. Your system should take an image with a number combination as input and return the interpreted text as numbers. The system must be called by the following syntax:

>>s=OCR(Image);

where Image is the matrix that is obtain when the following command is used

 >>[Image,Facit]=ReadNumberImage(P,N); 

and s is a string

The complete assignment specification

The complete assignment specification is found here.

Assistance

Christina Olsén (colsen@cs.umu.se) is responsible for the assignment

Images and Matlab-functions

Images for the assignment are found here. A zip-file, which includes all images, is found here.
The Matlab-functions: OCR_test and ReadNumberImage