We train various neural network architectures on the MNIST and HASY dataset and use image processing to read in handwritten mathetamical equations and expressions.
The project is broken up into 3 parts:
- Train an MLP to classify handwritten digits from the MNIST dataset and specific characters from the HASY database
- Extend the network to identify characaters and digits from images containing handwritten equations
- Evaluate mathematical expressions from the network classifications