Brain Tumor Segmentation and Model Optimization for 3-D Printing

Abstract

Medical image acquisition is generally done through Magnetic Resonance Imaging (MRI) or computerized tomography (CT) scan which is a non-intrusive technique. For identifying the Tumor in the brain, a segmentation process is used which involves using image processing algorithms on MRI scans. While many approaches have been used for brain tumor segmentation, the manual segmentation process is tedious and has inherent problems associated with it. This paper proposes an implementation of3-D slicer segmentation of the brain tumor model along with STL model optimization using Autodesk Netfabb. This will provide an effective way of reducing the STL errors in the segmented 3-D models. These optimized models can then be 3-D printed which will provide accurate information on tumor shape and size to the medical professionals. The process of STL model optimization shows very promising results on the MRI dataset used in this case study and it achieves the creation of error-free STL models for 3-D printing. Additionally, this work demonstrates the effective use of the 3-D slicer along with Autodesk NetFabb, to considerably reduce the STL errors which were generated in the conventional process of manual segmentation. The manual segmentation process was tedious and required force smoothening of 3-D slicer models obtained after segmentation leading to faulty 3-D model printing.