基于快速二维经验模态分解的纹理分割

Texture segmentation based on fast bidimensional empirical mode decomposition

经验模态分解是一种数据驱动的信号分解方法,具有局部性和瞬时性等特性,非常适合非稳态非线性信号分析。提出了一种新的快速二维经验模态分解方法,在新方法中,采用了新的边界抑制算法,改进了经验模态分解算法的筛选条件。将该方法应用于纹理分割,取得了满意的实验效果。

As a data-driven method for signal decomposition, empirical mode decomposition （EMD）, with the characteristics of partiality and instantaneity, is especially suitable for the analysis of the unstable and non-linear signals. A new and effective method for the bidimensional empirical mode decomposition is proposed. A new way for boundary restraint is introduced and the filter conditions in empirical mode decomposition is improved. Meanwhile, this new method has been applied to the texture segmentation, which has turned out to be effective and efficient according to the experimental results.