摘要
该文设计了一种基于支持向量机(support vector machine,SVM)的、关于MRI脑肿瘤图像的分割和评估系统,以便通过计算机辅助医疗来提高脑肿瘤的诊断和治疗过程中的效率和效果。该系统整合了一种新的特征选择方法(核空间类离散度),处理MRI图像序列所包含的三维数据信息并评估肿瘤体积的变化。整个系统实现了半自动运行,处理过程仅仅只有一次人工参与。通过实际病例上的实验验证、结果的定量评估和方法的比较,证明了该系统的有效性。
A system was developed for brain tumor segmentation and evaluation in magnetic resonance images(MRI) based on support vector machines(SVMs).The objective of this system was to improve the effectiveness and efficiency of computer aided diagnosis.3-D information from MRI sequences is utilized through a feature selection method and kernel class separability(KCS) to evaluate the tumor volume change.This system is semi-automatic with very simple operator interaction.Comparison among several methods using real patient data demonstrate the system effectiveness.
出处
《清华大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2011年第1期96-100,共5页
Journal of Tsinghua University(Science and Technology)
关键词
图像分割
脑肿瘤
三维信息
支持向量机
image segmentation
brain tumor
3-D information
support vector machine(SVM)