This project built upon an image stylization algorithm that uses highly irregular regions to abstract the image. These regions still respect image edges. Images are sent through a process where its pixels separated into these regions. After these regions are created, they are coloured using the average pixel colour in the region. A black and white image abstraction was also desired, which required an error diffusion algorithm to be implemented into this algorithm. This project investigated ways to implement a contrast-aware error diffusion algorithm to achieve this. The project began with a simplistic version of the error diffusion algorithm that assumed all regions were equivalent sizes before progressing to an implementation that considers the size of the regions. The final implementation of this algorithm was the contrast aware implementation. When this implementation used a basic error calculation the results were quite successful abstractions of the input image. This was also dependant on the input image as some inputs can give poor results depending on their intensity level. The results under more complex error calculations were not as successful but many tests and investigations were done to determine why this was happening. The results from these tests will also be included.