This is a prototype where you can draw your shorthand character to translate to text
Teeline classification undergoes the process of a 3-Dimensional Convolution Neural Network (CNN) trained on a dataset Teeline alphabetical characters.
The most challenging part (usually) to any machine learning task, is creating the dataset. Especially with such a small database of Treeline transcriptions available to us. We had to overcombe this challenge with a pre-processing stage for data cleaning and then further heavily augmenting the dataset to create a reasonable sample size for the CNN.