Accelerated parallel texture optimization

Texture optimization is a texture synthesis method thatcan efficiently reproduce various features of exemplar textures.However, its slow synthesis speed limits its usage in many interactiveor real time applications. In this paper, we propose a parallel textureoptimization algorithm to run on GPUs. In our algorithm, k-coherencesearch and principle component analysis (PCA) are used for hardwareacceleration, and two acceleration techniques are further developed tospeed up our GPU-based texture optimization. With a reasonableprecomputation cost, the online synthesis speed of our algorithm is4000+ times faster than that of the original texture optimizationalgorithm and thus our algorithm is capable of interactiveapplications. The advantages of the new scheme are demonstrated byapplying it to interactive editing of flow-guided synthesis.