It is ubiquitous that meaningful structures are appear over textured surfaces. Extracting them un-der the complication of texture patterns is very challenging, but of great practical importance. Consequently, we have proposed a novel structure-aware filter via bilateral kernel regression with a variational struc-ture-kernel descriptor which can extract main structures from textures through variational structure-kernel descriptor and incorporate its results into the bilateral kernel regression: we first apply the related reductive texture decomposition to construct the structure-kernel descriptor. Then, we incorporate the descriptor into the bilateral kernel regression to achieve an expected structure-preserving output. Algorithmically, we pro-pose a numerically stable approximation iterative procedure to achieve effective implementation. At last, some experimental results are presented to demonstrate that our approach leads to better or comparable per-formance and is effect in some applications, such as HDR, super-pixel segmentation and so on.%图像中结构被纹理覆盖的现象无处不在，将结构从复杂的纹理中提取出来是一件非常有挑战但又极其重要的事情。为此，提出基于双边核回归的纹理分解方法，通过局部全变分的核描述子能将结构和纹理很好地区分开来，并将结果与双边核回归框架相结合：采用相对约减纹理分解来构造结构核描述子；再将该描述子与双边核回归融合来获得期望的结构感知滤波输出；最后采用一个稳定近似迭代的流程来实现所提算法。实验证明，与其他边缘感知滤波方法相比，文中方法能获得更佳的结构保持效果，并能被推广到多个视觉相关的应用上，如高动态范围色调映射、超像素分割等。
展开▼
