Example-based Texture Synthesis

Abstract

Textures have traditionally been used in computer graphics to enhance the appearance of scenes, e.g., texture mapping provides an inexpensive way of representing surface detail. However, acquiring the texture in the first place can be a challenging problem. Example-based texture synthesis is a way of solving this problem by generating a high resolution texture from a small input example. While image texture synthesis is the classic problem in the field, this idea has also been extended to other domains like video, geometry, surface textures, and more generally, texture functions, which describe a varying texture as a function of time, lighting/viewing direction, or some other such variable(s). Texture has also been studied in the context of image understanding and analysis, and research in this area has often provided insight for texture synthesis algorithms. Not surprisingly, texture synthesis finds applications in problems related to image/video editing, merging, and completion. More recently, a lot of research has also focussed on improving the controllability and interactivity of texture synthesis techniques, which promises to broaden their applicability. This course will examine some of the recent advances in texture synthesis algorithms and discuss them in the context of different applications. We will present various algorithmic paradigms for texture synthesis including statistical, non-parametric, as well as optimization-based techniques. We will discuss issues related to representation and processing of textures across different modalities and applications such as images, video, geometry, surface textures, globally variant texture-functions, interactive synthesis on the GPU, and fluid texturing.

Who should attend?

Prerequisites