Improved Automatic Generation of Curved Nonograms

Abstract

In this thesis we propose a new algorithm for automatically generating curved nonograms, which are a variation on the popular nonogram puzzle. Aside from a core algorithm, several optional features are introduced which can modify the algorithm to try and generate curved nonograms that best fulfill the aesthetic requirements for publishing.The core algorithm and additional features are tested on several different input images. We find that the algorithm is able to generate usable puzzles in many cases. Some of the optional features prove to be promising, but require further study and adjustment to get their desired effect.