一种改进的边界法向量叠加疑似肺结节提取-JournalofNortheastern.PDFVIP

第31 卷第8期 东 北 大 学 学 报 ( 自 然 科 学 版 ) Vol131, No. 8 2 0 10 年 8 月 Journal of Northeastern U ni ersity( Natural Science) Aug. 2 0 1 0 栾国欣, 魏 颖, 薛定宇 ( , 110004) : , , ROI ; # ROI , , , ROI ; , , ; # , , # # : ; ; ; ; : TP 391. 4 1 : A : 1005-3026( 20 10) 08- 1078-04 An Improved Algorithm Based on Boundary Normal Vector Overlap for Suspected Pulmonary Nodule Extraction L UAN Guo-xin, WEI Ying, X UE Ding-yu ( School of Information Science Engineering , Northeastern Uni ersity, Shenyang 110004, China. Correspondent: LUA N Guo-x in, E-mail: luanguoxin @ 163. com) Abstract: An impro ed boundary normal ector o erlap algorithm based on local shape constraint and adapti ely defined normal ector magnitude is proposed to detect suspected pulmonary nodules. The original region of interest( ROI) is segmented w ith adapti e thresholding method in pulmonary parenchyma, then the local con ex-conca e shape feature is gained for e ery pixel of initial ROI boundaries to satisfy the calculations of normal ector direction and adapti ely defined normal ector magnitude of e ery con ex pixel on the initial ROI boundaries. O erlapping the normal ectors, the locally maximum o erlap can be chosen so as to detect the suspected round pulmonary nodules of different sizes. The restriction due to the local con ex-conca e shape feature should be taken into account before o erlapping the normal ectors, thus simplifying the calculation of o erlap. The adapti ely defined normal ector magnitude is a ailable to get rid of the limitation of fixed-size nodules as the result of detection. Experiment resul