Constrained Graphcut Texture Synthesis

Abstract

This paper describes constrained graphcut texture synthesis (CGS), a graphcut-based synthesis algorithm that creates output textures satisfying constraints. We show that constrained texture synthesis can be posed in a principled way as an optimization problem that requires balancing two measures of quality: constraint satisfaction and texture seamlessness. We then present an efficient algorithm for finding good solutions to this problem, using generalized graphcut minimization. CGS enables explicit control while preserving the speed and quality benefits of graphcut texture synthesis. This approach supports the full image analogies framework, while providing superior image quality and performance. A range of applications of CGS are demonstrated, including detail synthesis, artistic filtering by analogy, and texture-by-numbers. CGS is easily extended to handle multiple constraints on a single output, thus enabling novel applications that combine both user-specified and image-based control.