Handwritten Hangul OCR

• 

  Testing model on real data

Inspiration

Korean language

What it does

It takes an image of a Korean character as an input and classifies it by using SVM classifier.

How we built it

I have used SVM to train the model and have used accuracy and confusion matrix to evaluate the model.

Challenges we ran into

Depending upon the train and test data, the accuracy of the model changes.

Accomplishments that we're proud of

Current accuracy of model - 92.22%

What we learned

Using image dataset and creating proper data

What's next for Handwritten Hangul OCR

Improving the model accuracy

Built With

• python