readme.txt

this is digit recognition using som

A neural network has been trained using matlab to recognize digits as a CMPT 310 project in Simon Fraser University (BC, Canada).  This is a group project done by Vincent Chu and Kay Yang.  The script included can be used to recognize digits (0 to 9 arabic characters) images at a 90% accuracy.  The target images being loaded to matlab must be binary and must 
have size 28x28.

The script included can only recognize digit.  To train the neural network (and to obtain the weights included in the script), a separate script was used.  

For more information, feel free to e-mail me at:
chuvincent (at) gmail(dot) com

http://www.sfu.ca/~vwchu

How to use the script
=====================

Once you extracted all the files in the zip file, load digitrecognition.mat.  (If you're using a version of matlab that predates version 7, you will get a file corruption error.  You should load digitrecognitionv6.mat instead). It would load 10 test images, and 10 corresponding labels indicating what those images are. 

1. Type recognizeDigits(images(:,:,1)) to recognize the first image as an example.  In general, you can have recognizeDigits(images(:,:,i)), where i is an integer from 1 to 10.

2. labels(1) would show what the image supposed to be (human-recognition).  

In this sample, machine recognition of images 1-9 are correct.

Note:
a) To view the image, simply type figure; imshow(images(:,:,i)); where i is a number from 1 -10 (the index into the image array).
b) image i in images array does not correspond to digit i. The 10 sample images are chosen at random and contain random digits