Anisotropic dynamic-morphological-diffusion for color image segmentation
Junji Maeda, Takehiro Harada, Sato Saga, Yukinori Suzuki · 2012
In this paper, we present a segmentation method of noisy color images that uses an anisotropic diffusion algorithm and a region-growing algorithm. We propose a modified anisotropic diffusion algorithm as a new precise edge-preserving smoothing technique by using the dynamic morphological filtering. We incorporate a dynamic selection of multiple structuring elements for each pixel in the morphological filter before the anisotropic diffusion in order to improve the ability of edge-preserving smoothing. As a result, noise and unnecessary details of images are effectively smoothed while preserving small image structures before performing a region-growing algorithm. Therefore, the proposed method is suitable for an accurate segmentation of color images with noise. Several simulated examples are presented that demonstrate the effectiveness of the proposed technique.