HRCT image segmentation algorithm based on tolerance granular space model
Xinying Xu, Tianrui Cao, Chengdong Yan, Gang Xie, Zhifeng Wu · 2009
In order to resolve quantitative analysis of lung tissue, according to the features of the medicine CT image's complicated texture, a image segmentation approach based on region growing method and granular computing is presented in this paper. This segmentation algorithm is based on tolerance granular space model of granular computing. First, describes chest high-resolution CT images (HRCT) as granules and establishes the tolerance granular space model. Then, this algorithm can chooses an image sub-block as the seed block according to the intension of tolerance granule and carries on the region growing segmentation according to the tolerance relations automatically. Extensive experiments and evaluations were carried out and the results illustrate that this method can segment HRCT image accurately and precisely, and can obtain the lung tissue removing the tiny blood vessel and the trachea.