A Novel Fast Marching Segmentation Algorithm for Pulmonary Nodules in Chest Radiographs
Segmenting nodule region in chest radiographs is a particularly challenging problem due to the complexity and variability of human anatomy.In this work,in order to solve the boundary leakage problem in conventional fast marching method,we present a novel fast marching segmentation algorithm for pulmonary nodule in chest radiographs.The proposed algorithm begins by training a nodule model via support vector machine classification algorithm as a prior knowledge.Then it calculates the decision value on each pixel in chest image to constrain front marching in fast marching evaluation process.The performance of our method is evaluated on 50 nodule images set and compared with conventional fast marching method and Watershed method respectively.The preliminary results show that our proposed algorithm is superior to the above two methods distinctly and can raise the robustness and precision of nodule segmentation due to the benefit of nodule prior knowledge.
Fast Marching Support Vector Machine(SVM) Nodule Segmentation
Qiyong Guo Mantao Xu Jiwu Zhang
Department of Computer Science and Engineering,Fudan University,Shanghai,China Global R&D Center,Carestream Health Company,Shanghai,China
国际会议
7th Asian-Pacific Conference on Medical and Biological Engineering(第七届亚太地区生物工程学术会议)
北京
英文
2008-04-22（万方平台首次上网日期，不代表论文的发表时间）