基于水平集的三维脑部肿瘤分割_薛晓君.pdf

20 12 V ol. 20 N o. 12 20 10 12 CO M PU T ER T ECHN OL O GY AN D DEV EL OPM ENT Dec. 2010 1 1 2 1 薛晓君 , 张立强 , 薛 忠, 杨建华 ( 1. 西北工业大学 自动化学院, 陕西 西安7 10072; 2 . Methodist 医院研究所生物工程信息 心, 美国德克萨斯州休斯顿 77030) : , , , , , : ; ; ; : T P39 : A : 167 3- 629X( 2010) 12- 0201- 04 A Brain Tumor Segmentation Method Based on Level Set 1 1 2 1 X UE Xiao- un , ZHAN G L-i qiang , XUE Zhong , Y AN G Jian- hua ( 1. A utomatic School , N orthw est Poly technic U niversity, Xi an 710072 , China; 2 . Center for Bioeng ineering and Informatics, M ethodist Hospital Research Institute and Department of Radiolog y, M ethodist Hospital, Weil Cor nell M edical College, Houston 77030, U SA) Abstract:Brain t umor segment at ion is a very im portant image processing st ep in diagnosis, t reat ment and research. But it is st ill a chal- lenging t ask due t o vary ing in size, shape, locat ion, and image int ensit ies w it hin and around t he t umor . Proposed a new brain tumor seg- ment at ion met hod based on t he level set method using t he gradient and the int ensit y dist ribution in format ion in magnet ic resonance im- ages. A new energy funct ion is def ined to mat ch t he brain t umor int ensity dist ribut ion more accurat ely . T he proposed met hod is used t o brain magnet ic resonance images, it can segment t he t umor correct ly . T he proposed met hod segments correct ly compared w it h the t radi- t ional level set m ethod. Key words:image segmentation; level set; brain t umor; magnet ic resonance imaging 0