The aim of this project is to take an input gray level or colour image and generate from it a binary (i.e. black & white) output image. While this can be easily done with a simple thresholding approach, this project will investigate techniques that attempt to retain the essence of the input image and produce a more artistic binary image.

The topic of this project comes under the heading of non-photorealistic rendering - which combines aspects of image processing with computer graphics.

My chapter http://users.cs.cf.ac.uk/Paul.Rosin/resources/papers/npr-Yukun-book.pdf gives an idea of traditional approaches that you could consider as a starting point. Alternatively you could try deep learning approaches.