AN EFFICIENT SYNTHESIS APPROACH FOR PERSPECTIVE DISTORTED TEXTURES USING CONTROL STRUCTURE CORRECTION

摘要

Perspective distorted textures are pervasive in man-made and real world, which are usually captured by camera lens with non-vertical camera axis. However, the conventional patch-based synthesis approaches cannot be applied to the non-parallel textures. In this paper, we construct a control structure correction model based on an automatic rectification approach to synthesize the appearance of the perspective distorted textures faithfully. Furthermore, some common drawbacks of patch-based sampling algorithm such as blending procedure and high computational cost will be also addressed in this paper. In this paper, we develop a hybrid blending scheme to improve visual transition artifacts along patch boundaries, and also apply Particle Swarm Optimization to improve the computational efficiency. The experimental results and comparisons demonstrate that our proposed control structure correction through Hough transform work effectively and efficiently for the synthesis of perspective distorted textures.