A semi-automatic approach to the segmentation of liver parenchyma from 3D CT images with Extreme Learning Machine.

This paper presents a semi-automatic approach to segmentation of liver parenchyma from 3D computed tomography (CT) images. Specifically, liver segmentation is formalized as a pattern recognition problem, where a given voxel is to be assigned a correct label - either in a liver or a non-liver class. Each voxel is associated with a feature vector that describes image textures. Based on the generated features, an Extreme Learning Machine (ELM) classifier is employed to perform the voxel classification. Since preliminary voxel segmentation tends to be less accurate at the boundary, and there are other non-liver tissue voxels with similar texture characteristics as liver parenchyma, morphological smoothing and 3D level set refinement are applied to enhance the accuracy of segmentation. Our approach is validated on a set of CT data. The experiment shows that the proposed approach with ELM has the reasonably good performance for liver parenchyma segmentation. It demonstrates a comparable result in accuracy of classification but with a much faster training and classification speed compared with support vector machine (SVM).