基于主成份分析和支持向量机的MRI图像多目标分割

The Segmentation of Multi-target MRI Image Based on Principal Component Analysis and Support Vector Machine

在MRI图像中,颅内各组织的边界极其复杂且不规则,这对传统的分割算法提出严峻的挑战。主成份分析(PCA)可达到降维和消除冗余信息的目的,为使支持向量机(SVM)准备的样本空间更为紧凑、合理。本研究采用PCA将图像的57维特征向量处理后,研究多分类SVM对MRI图像进行多目标分割,成功提取颅内各组织不规则边界。理论和实验表明,基于PCA和SVM相结合的分割性能优于仅采用SVM的分割性能。

In MRI image, the boundary of each encephalic tissue is very complicated and irregular. It is a big challenge to the traditional segmentation algorithms. Principal Component Analysis （PCA） can realize dimension reduction and eliminate redundant information, make the sample space for Support Vector Machine （SVM） more compact and reasonable. In this paper, PCA was used to process 57 dimensional feature vectors of image and the segmentation of multi-target MRI image based on multi-classification SVM was investigated. The result showed that the irregular boundary of each eneephalic tissue was extracted successfully. The theory analysis and experimental results indicated that the segmentation based on PCA and SVM displayed better performance than the segmentation based on SVM only.