关键词: 分数阶微分, 增强, 分割, 最优阈值, 血管

Abstract: In order to improve the accuracy of automatic vascular segmentation in CT lung images, after researching image enhancement, segmentation method and the enhancing ability of the fractional differential operator toward image details, a local sub-regional segmentation method based on fractional differential enhancing was proposed. This method enhanced the lung CT images with fractional differential operator firstly, and then the optimal threshold acquired under two control indexes was used for vascular segmentation in the local region. The experimental results show that the proposed method can effectively extract the vascular network in which the blood vessels have more detailed information. And compared to traditional pulmonary vascular segmentation method, it has more accurate segmentation ability of the pulmonary vascular in CT images.

Key words: fractional differential, enhancement, segmentation, optimal threshold, vascular