METHOD FOR FORMING 3D MAXILLOFACIAL MODEL BY AUTOMATICALLY SEGMENTING MEDICAL IMAGE, AUTOMATIC IMAGE SEGMENTATION AND MODEL FORMATION SERVER PERFORMING THE SAME, AND STORAGE MEDIUM STORING THE SAME

摘要

Disclosed is a method for forming a three-dimensional (3D) model of skin and mandible by automatic medical image segmentation which is performed in an automatic image segmentation and model formation server. The method includes (a) receiving 3D medical image data that is a set of two-dimensional (2D) images for horizontal planes of a face, (b) obtaining a contrast histogram based on distribution of contrasts of the 3D medical image data, and segmenting the 3D medical image data for the face into multiple regions separated into at least one partial region based on the contrast histogram, (c) extracting only the face by removing portions other than the face from the multiple regions for the face, and extracting a skin region of the face, (d) extracting the mandible from each of the 2D images for the horizontal planes of the face through a 2D detailed segmentation technique using an active contour method based on a level set function, and (e) reconstructing the extracted skin region and mandible as the 3D model.

主权项

1. A method for forming a three-dimensional (3D) model of skin and mandible by automatic medical image segmentation which is performed in an automatic image segmentation and model formation server, the method comprising:
(a) receiving 3D medical image data that is a set of two-dimensional (2D) images for horizontal planes of a face; (b) obtaining a contrast histogram based on distribution of contrasts of the 3D medical image data, and segmenting the 3D medical image data for the face into multiple regions separated into at least one partial region based on the contrast histogram; (c) extracting only the face by removing portions other than the face from the multiple regions for the face, and extracting a skin region of the face; (d) extracting the mandible from each of the 2D images for the horizontal planes of the face through a 2D detailed segmentation technique using an active contour method based on a level set function; and (e) reconstructing the extracted skin region and mandible as the 3D model.