EECS 349, Machine Learning

  Insturctor: Prof. Doug Downey
By Majed Valad Beigi
majed.beigi@northwestern.edu
Northwestern University
home Introduction Design Back Propagation NN LAMSTAR NN SVM Results Conclusion

Offline HandWritten Character Recognition
 

You can download the full project report and source code here:
FullReport_Source Code

OVERVIEW

An off-line handwritten alphabetical character recognition system using Back Propagation Neural Network (BP NN), LAMSTAR Neural Network (LAMSTAR NN) and Support Vector Machine (SVM) is described in this project. The general steps of the algorithm are: 1- Scanning the source material (a paper with all the characters written on it) using an optical scanner, 2- Performing automatic processing on the image, 3- Creating the input dataset for Artificial Neural Network (ANN) or SVM by extracting the most important attributes from the image of each character and representing the features in the form of a matrix of ‘0’s and ‘1’s(attributes are important and can have a crucial impact on end results) 4- Classifying the dataset using BP NN, or LAMSTAR NN or SVM, and performing Recognition during the test 5- Finally, getting the results of the recognition. In the next parts of the website I will explain each part in more details.

For the results, four data sets, each containing 52 alphabets (26 Upper-Case and 26 Lower-Case characters) written by various people, are used for training the neural network and 520 (10 per each 52 characters) different handwritten alphabetical characters are used for testing. The proposed recognition system performs quite well. Experimental results show that the LAMSTAR NN with the success rate of 93.84% and the training time of 1.2 Seconds is the fastest and the most efficient compared to BPNN and SVM for this problem.