Handwritten Digit Classifiation using support vector machines.
Dependencies used:
OpenCV2
Scipy
ScikitLearn/SKlearn
Numpy
and some Standard libraries.
"generateClassifier.py" generates a recognition model after training.
"performRecognition.py" is used to classify the digits.(Takes two agrguments "-classifier" and "-i"(image)).
"mnist-original.mat" contains all the training data.
"deploy.bat" can be used to deploy the model.(already contains the path to the classifier,just need to add the argument for image path) Rest are example images.
Usage:
Use "python performRecognition.py -classifier(PATH TO CLASSIFIER) -i(PATH TO TEST IMAGE)
Might throw out some weird errors at first,properly edited images with a singular color background is advised to be used.