This paper proposes a new algorithm for image decomposition, termed bidimensional ensemble patch transform that adopts a patch process and an ensemble process. The proposed method is based on a nontrivial two-dimensional extension of the one-dimensional ensemble patch transform. It is required to have a successful recognition of local characteristics of the image. The proposed method is capable of efficiently decomposing an image into two-dimensional intrinsic mode functions considering spatial patterns and yields decomposition results that are robust to random fluctuations of the image such as noises. Furthermore, the proposed method is computationally easy to implement.

中文翻译：

---

二维整体补丁变换的图像分解

本文提出了一种新的图像分解算法，称为二维整体补丁变换，该算法采用了补丁过程和集合过程。所提出的方法基于一维整体补丁变换的非平凡二维扩展。需要成功识别图像的局部特征。所提出的方法能够有效地将图像分解为考虑空间模式的二维固有模式函数，并产生对图像的随机波动（例如噪声）具有鲁棒性的分解结果。此外，所提出的方法在计算上易于实现。