An Improved Method for Building A 3D Model from 2D DICOM

Abstract

The development of images processing techniques in recent years brings us a lot of useful application in practice. In the field of medical image processing, several researches and their application have been widely applied in practice such as diagnostics, treatment, training and education. This paper presents an improved method for reconstructing the shape of 3D objects from 2D DICOM datasets. We first focus on DICOM data processing based on techniques of digital image processing to get a 3D point cloud. This step includes extracting features, boundary of the shape, removing noisy data and inserting new pixels to obtain a regular dataset on each 2D DICOM slice. In the next step, we reconstruct the 3D object based on triangulation of 3D point clouds. In the last step, we build an application for rendering and visualizing the 3D medical image objects. Comparing to the existing methods, the proposed method is very fast for creating a 3D object from DICOM images. The application can be used to support medical staffs and doctors in diagnostic, treatment and training.