Marching cubes: A high resolution 3D surface construction algorithm
SIGGRAPH '87 Proceedings of the 14th annual conference on Computer graphics and interactive techniques
Scale-Space and Edge Detection Using Anisotropic Diffusion
IEEE Transactions on Pattern Analysis and Machine Intelligence
Parametrization of closed surfaces for 3-D shape description
Computer Vision and Image Understanding
The nature of statistical learning theory
The nature of statistical learning theory
A Tutorial on Support Vector Machines for Pattern Recognition
Data Mining and Knowledge Discovery
Small Sample Size Learning for Shape Analysis of Anatomical Structures
MICCAI '00 Proceedings of the Third International Conference on Medical Image Computing and Computer-Assisted Intervention
An introduction to variable and feature selection
The Journal of Machine Learning Research
Ellipsoidal support vector clustering for functional MRI analysis
Pattern Recognition
| Hi-index | 0.01 |
Statistical morphometric analysis is used in this paper to assess the skull morphological difference in adolescent idiopathic scoliosis (AIS) patients as compared with normal controls. The skull morphology analysis starts from a 3-D segmentation of the skull in MRI data using the adaptive thresholding and mathematical morphological operations. The skull vault is then separated from the whole skull surface using a demarcation plane determined by three anthropometrically meaningful landmarks. Since the segmented skull surface is not sufficiently smooth, which will introduce errors in the subsequent shape description, we devise a surface smoothing method to improve the quality of the skull surface. This mesh smoothing task is formulated as a quadratic optimization problem, in which the objective function is designed to be a linear combination of the shape term and the smoothing term, thus it overcomes the shrinkage problem in the traditional Laplacian smoothing. A spherically uniform sampling on the smoothed skull surface is performed afterwards. The scale-normalized distances from the skull vault centroid to the sample points are defined as features, on which the recursive feature elimination strategy is applied to select the features that contribute most to the classification of AIS patients and normal controls using the support vector machine. The results of this study quantify the skull vault surface changes and provide new evidence of bone formation abnormality in AIS patients.