Feature extraction of handwritten characters

I am doing my final year project on "Hand written character recognition" and I am not getting how to extract features of characters. Can anyone please post the sample code for extracting features of an image containing a character?

Answer 1

[Shashank Prasanna]

What features are you looking for? There are many features you can extract from an image.

Here is an example of hand writing recognition that simply uses the data as it is:

http://www.mathworks.com/machine-learning/examples.html?file=/products/demos/machine-learning/handwriting_recognition/handwriting_recognition.html

[dream]

I just want about 8 to 9 features so that the feature vector can be fed into neural network. Can you suggest a sample code?

[Shashank Prasanna]

To answer your question you'll have to tell us what features you want. If you don't tell me what those 8, 9 features are, I'm afraid you won't get much help. Take a look at Image Analyst's post and hopefully that may inspire you to take the next step in your project. Best of luck!

[dream]

@shashank That is the problem I am facing. I don't know what features to extract. I am not getting much help from the internet either. So please help

[Image Analyst]

I find it hard to believe that none of the papers that did OCR on handwriting said what features to extract. Did you actually look at any of them?

Answer 2

[Image Analyst]

There are plenty of published handwriting algorithms here: http://iris.usc.edu/Vision-Notes/bibliography/contentschar.html#OCR,%20Document%20Analysis%20and%20Character%20Recognition%20Systems. Pick one and code it up.

Here is an example how how to measure various things (area, perimeter, etc.) in an image (of coins): http://www.mathworks.com/matlabcentral/fileexchange/25157-image-segmentation-tutorial-blobsdemo