摘要： 基于径向基函数的分块二维经验模态分解算法在拼接处的不连续失真较严重。为此，提出一种改进的分块二维经验模态分解算 法。通过计算相邻图像块的最小对称相似性，确定相邻图像块间的重叠尺寸，利用自适应无缝拼接算法，对图像分解结果进行融合。实验结果表明，该算法能较好地改善拼接失真现象。

关键词: 分块二维经验模态分解, 对称相似性方差, 自适应拼接, 径向基函数

Abstract: The blocking artifact is noticeable in the block partitioning Bidimensional Empirical Mode Decomposition(BEMD) based radial basis function interpolation. For alleviating the blocking artifact problem, it proposes a modified blocking BEMD method. This paper calculates the enlargement size of each image block by the minimal symmetry resemblance variance of adjacent blocks, and the decomposing results of each block are fused by an adaptive seamless splicing method. Experimental results show that, this algorithm can improve the phenomenon of montage distortion.

Key words: blocked Bidimensional Empirical Mode Decomposition(BEMD), symmetry resemblance variance, self-adaptive splicing, radial basis function

中图分类号: 

• TP301.6