Optical Recognition of Handwritten Digits

Hello, i need to develop a sistem for the automatic recognition of digits from written data (8x8 images). Specifically i need to determine which couple of digits is harder to discriminate and why.

I have a problem in the import of data in the workspace, the Dataset is in this link, http://archive.ics.uci.edu/ml/datasets/Optical+Recognition+of+Handwritten+Digits

I don't understand how to load the training set (optdigits-origin.tra) in my workplace, in class we usuall use readtable but i don't think it's very useful here, my professor said we should use load but if i do that i get an error, can anybody please help?

Thanks

Answer 1

[Shashank Gupta]

You can directly use the load function to load the data in the workspace. I try loading it and able to properly load it in the workspace using the same function.

I just used.

% loading the data to workspace.
load('optdigits.tra');

Can you verify if you are able to do it. If you already tried and getting some error, can you also paste the error here so that I can investigate.

Cheers.

[Fabrizio Lepori]

Yes thank you, I used the function load and is now working.

in order to show the digits i used this code, can you tell me if you think is correct?

Y = load('optdigits.tra');     %load training set 
figure
imagesc(reshape(Y(1,1:64),8,8)');    %show the first image
axis square;
colormap(gray);
X = load('optdogits.tes')     %load test set

now i'd like to show the 10 digits in a subplot using the test set, do you think a for loop would be the best solution?

Thank you for the help so far.