Hand Writing Recognition Term Project


ECE597B -- Neural Network Design



MATLAB

These are some M files that I have written for the class. Please email me if you find any errors
Learning Rate (Chapter 9)LMS (Chapter 10)

HP48G(X) Neural Network Programs


LMS - Least Mean Square PINV - Pseudo-inverse of a matrix
MLST - Matrix to List, used in LMS LRN - Learnrate, minimizing along a line (CHP 9)
XR-> - Vector to 1 row matrix, used in MLST

Finally Here!

Intructions on use:
1. Please be patient when the applet start, and when the applet is "Learning."
2. The left grid is the input grid, the right is the output pattern after calculated using the respective algorithm.
3. Draw figures and press Learn! Enjoy!

What does it do?

What it does is recognize "hand" or "mouse" written digits. It has two grids on the top. The left grid is the input grid, where you click using the mouse, to draw different numbers. The right grid is the output grid, where the program displays what numbers it thinks it is. To recognize click on the "Recognize" button. It'll apply the algorithm that's on the "choice bar." To learn a pattern: Draw a pattern of the input grid, then chose which rule or rules to learn the pattern using the check boxes. Wait until on the button right is says DONE. Then try recognizing...

For more information read my term paper it




const-javaguy.gif - 5.8 K One more semester....