I want to have a program that writes like a human. But I don't just want a font, but instead an 'intelligent' program that produce different result and that can be trained with different sets to generate different handwritings. As a training set I would like to have parts of a handwritten text (saved as a list of paths (like in vector graphics). Maybe as a means to simplify things, I could flatten the paths in to consecutive straight lines. My program receives a string of text and produces a list of paths (or a vector graphic, whatever is easier to work with)

My question now is: What kind of machine learning would be best to achieve this?

  • $\begingroup$ Welcome to Artificial Intelligence Stack Exchange! Questions about machine learning implementation are off-topic, as seen in the help center. $\endgroup$ – Matthew Graves Oct 3 '16 at 15:26

From the abstract of this paper Generating Sequences With Recurrent Neural Networks[pdf]

This paper shows how Long Short-term Memory recurrent neural networks can be used to generate complex sequences with long-range structure, simply by predicting one data point at a time. The approach is demonstrated for text (where the data are discrete) and online handwriting (where the data are real-valued). It is then extended to handwriting synthesis by allowing the network to condition its predictions on a text sequence. The resulting system is able to generate highly realistic cursive handwriting in a wide variety of styles.

An implementation is here: handwriting-synthesis

And using TensorFlow: Handwriting Generation Demo in TensorFlow

| improve this answer | |

Not the answer you're looking for? Browse other questions tagged or ask your own question.