Segmentation and Visualization of 3D Medical Images through Haptic Rendering

High-dimensional and high-resolution image data is increasingly produced by modern medical imaging equipment. As a consequence, the need for efficient interactive tools for segmentation and visualization of these medical images is also increasing. Existing software include state-of-the-art algorithms, but in most cases the interaction part is limited to 2D mouse/keyboard, despite the tasks being highly 3D oriented. This project involves interactive medical image visualization and segmentation, where true 3D interaction is obtained with stereo graphics and haptic feedback. Well-known image segmentation algorithms, e.g., fast marching, fuzzy connectedness, deformable models, and live-wire, have been implemented in a framework allowing the user to interact with the algorithms and the volumetric data in an efficient manner. The data is visualized via multi-planar reformatting, surface rendering, and hardware-accelerated volume rendering. We present a case study where liver segmentation is performed in CT images with high accuracy and precision.