Computerized Detection of Lung Nodules in CT Images by use of Multiscale filters and Geometrical constraint Region Growing
It is important to enhance and detect pulmonarynodules in computed tomography(CT) images in order to assistradiologists in the detection of lung cancer. Nodules which areincluded in medical image generally have multiple size and scaleand have blob-like structure. Recently, 3D multiscale filterapproach is proposed for lung nodules detection. However, the3D method takes too much computing time to be applicable incomputer-aided diagnostic (CAD). In this paper, we propose athree step method for lung nodule detection in CT images.Firstly, we use 2D multiscale filter to detect the candidates oflung nodules on the slice images. Secondly, distinguish nodulesbetween non-nodules which have blob-like shape by use ofgeometrical constraint region growing. Finally, extract shapefeatures of each region and use an automated rule-basedclassifier for reduction of false positives. The experimental resultsindicate that our algorithm can achieve a relatively highperformance for nodule detection.
Pei Xiaomin Guo Hongyu Dai Jianping
Sino-Dutch Biomedical and Information Engineering School Northeastern University Shenyang, China Department of Neurology Beijing Tiantan Hospital Capital Medical University Beijing, China
国际会议
成都
英文
1-4
2010-06-18（万方平台首次上网日期，不代表论文的发表时间）