Feature reconstruction for 3D medical images processing

Abstract

Computed tomography (CT) greatly improves the image information available to surgeons and is quickly becoming one of the most important methods to identify defects in patients. However, doctors must check these CT slices again and again to find the exact position of the problem. This is not only a waste of time but also highly inefficient. This paper presents a rapid process to reconstruct three-dimensional (3D) surfaces from two-dimensional (2D) CT scanned data. Based on geometric model reconstruction, a series of 2D CT image slices were read and transformed into profiles and 3D B-spline patch models were then generated. The data could be converted to a VRML format for virtual surgical applications. An actual rapid prototyping model was created for checking the design process. CT images of a human pelvic bone and skull mask were used as case studies to construct a 3D solid model from medical images. The actual model will allow doctors to observe CT slices in 3D, decrease training time, and provide references for medical image research.