Automatic liver segmentation in CT images based on Support Vector Machine
Jie Lu, Defeng Wang, Lin Shi, Pheng‐Ann Heng · 2012
Accurate and fully automated segmentation of liver parenchyma in medical images is necessary prerequisites for a variety of clinical and research applications, such as constructing three dimension anatomical model. In this paper, an automatic liver segmentation method based on Support Vector Machines (SVM) has been proposed. Segmentation is started by wavelet transform for image feature extraction. Subsequently, SVM is applied on the feature vectors for training and testing to realize pixel classification. Finally, region-growing is used to refine the result of SVM. Experiments have been conducted on different training-test partitions of the CT image datasets. Compared to manual segmentation provided by medical experts, our experimental results demonstrated the effectiveness of the proposed method.