关键词: 纹理分割, 方向分形维, 灰度特征, 模糊C均值聚类

Abstract: A new texture segmentation method based on directional fractal values and gray-level features is proposed in this paper.Firstly，the directional fractal dimensions and corresponding intercepts are extracted from a power spectrum image with a given local window，and their respective statistics（mean and variance） are combined with gray-level mean and variance to form a multidimensional feature vector.Then we apply the fuzzy c-means algorithm as the classifier based on formed feature vectors to get the texture segmentation result.The experiment shows that the proposed method is significantly effective and robust both on textile and medical images.

Key words: texture segmentation, directional fractal dimension, gray-level feature, fuzzy C-means