Simple Deep Neural Network to Classify 43 different classes of Mathematical Characters(CROHME Dataset) Using this repository https://github.com/ThomasLech/CROHME_extractor to extract the data into train and test folders and then running the data through simple deep neural network to to classify between the 43 classes.
The approximate VALIDATION accuracy was : 82 %
This is not an optimum solution as it uses a naive deep network (commonly use for MNIST), further i will imply the same using a convnet which will be a more efficient solution to this problem