Local resampling for patch-based texture synthesis in vector fields

Abstract

In order to synthesize distortionless texture on surfaces, we develop a direct and accurate approach for local resampling in vector fields, and then use the approach to synthesize textures on 2-D manifold surfaces directly from a texture exemplar. Regular-grid patches produced by the local resampling are used as building blocks for texture synthesis. Then texture optimization and patch-based sampling are generalized to synthesize texture directly in vector fields. The first scheme can create texture of higher quality; however the second scheme is faster and simpler and works well for a wide range of textures. Users can control the vector field on the mesh to generate textures with local variations including the orientation and scale. Many experimental results are presented to demonstrate the ease of use and reliable results provided by our system.

Discrete texture elements synthesis on surfaces using elements distribution

“…As Figure 3 e , e tar ) over the surface is not equivalent to the length of edge over 2D plane. Inspired by mesh resampling (Chen et al, 2010), we formulate the computation of the position of new seeds as an approximate evaluation problem in both steps. First, we sample a number of candidate points on each face of the coverage range of the neighbourhood ω(e tar ).…”

“…Texture synthesis on surfaces (Wei et al, 2009) has proved to be a powerful technique for reproducing a similar texture that fits an arbitrary surface naturally using a given 2D example texture. It has now reached an advanced degree of maturity due to lots of work (Chen and Dong, 2008;Chen et al, 2010;Kwatra et al, 2007;Lefebvre et al, 2005;Magda and Kriegman, 2003;Praun et al, 2000;Soler et al, 2002;Tong et al, 2002;Turk, 2001;Wei and Levoy, 2001;Zhang et al, 2003) towards synthesising textures directly on surfaces. Although most previous approaches (Chen and Dong, 2008;Chen et al, 2010;Kwatra et al, 2007;Praun et al, 2000;Soler et al, 2002) work well for a wide range of textures, which embody the simple idea of texturing a surface with overlapping patches, they fail to produce good results for many textures spatially with repeated texture elements.…”

Feature-aligned shape texturing

“…[2000], where user-specified irregular texture patches are randomly pasted onto a surface. Texture optimization [Kwatra et al 2005] has recently been extended to surfaces Chen et al 2009], multiscale texture synthesis from multiple texture exemplars and inverse texture synthesis [Wei et al 2008]. Fu and Lung [2005] have extended Wang tiles ] to surfaces.…”

“…Vector field guided synthesis has been previously achieved in an optimization-based framework in both 2D [Kwatra et al 2005] and 3D Chen et al 2009]. We observe that while the result of such un-constrained optimization conforms to the overall surface shape, as in Figure 2(c), when the texture exemplar contains prominent feature lines, such as those in Figure 5(a), being able to align these texture features exactly to surface features would greatly enhance the shape appearance, as can be seen in Figure 2(d).…”

Feature-aligned shape texturing

“…Therefore, we have to ensure that the distance between neighboring sample vertices in † is less than . We follow the sampling method in [Chen et al 2009] to obtain † . Specifically, we start by uniformly sampling the vertices using distance along the salient curves, where is a user-chosen parameter in [0, 1] to control the sample spacing.…”