Research Projects

Abstract. Topology-preserving geometric deformable models (TGDMs) are used to segment objects that have a known topology. Their accuracy is inherently limited, however, by the resolution of the underlying computational grid. Although this can be overcome by using fine-resolution grids, both the computational cost and the size of the resulting surface increase dramatically. In order to maintain computational efficiency and to keep the surface mesh size manageable, we have developed a new framework, termed OTGDMs, for topology-preserving geometric deformable models on balanced octree grids (BOGs). In order to do this, definitions and concepts from digital topology on regular grids were extended to BOGs so that characterization of simple points could be made. Other issues critical to the implementation of OTGDMs are also addressed. We demonstrate the performance of the proposed method using both mathematical phantoms and real medical images.

Octree-based Topology-preserving Isosurface Simplification

Abstract. Isosurface generation has many important applications in medical imaging. Standard isosurface algorithms generate very large triangle meshes when high resolution volumetric data is available, which increases rendering time and storage requirements. Most existing mesh simplification algorithms either do not guarantee non-intersecting meshes or require large cost to prevent self-intersection. We present an octree-based isosurface generation and simplification method that preserves topology, guarantees no selfintersections, and generates a surface that approximates the true isosurface of the underlying data. Rather than focusing on directly simplifying the surface mesh, the new strategy is to generate an octree grid from the original volumetric grid in a way that guarantees these desired properties of the generated isosurface. The new method demonstrates savings of 70% in mesh nodes for real 3D medical data with highly complicated shapes such as the human brain cortex and the pelvis. The simplified surface stays within a userspecified distance bound from the original finest resolution surface, preserves the original topology and has no selfintersections.

Super-resolved Multi-channel Fuzzy Segmentation of MR Brain Images

Abstract. We propose a new fuzzy segmentation framework that incorporates the idea of super-resolution image reconstruction. The new framework is designed to segment data sets comprised of orthogonally acquired magnetic resonance (MR) images by taking into account their different system point spread functions. Formulating the reconstruction within the segmentation framework improves its robustness and stability, and makes it possible to incorporate multispectral scans that possess different contrast properties into the super-resolution reconstruction process. Our method has been tested on simulated and real 3D MR brain data.

Superresolution Reconstrcution of MR Brain Images

Abstract. MR images typically have poorer resolution in the slice-selection direction than the in-plane directions. In this work, we adopt an MAP super-resolution method to reconstruct a high-resolution image from two orthogonal scans of the same subject. The new image shows improved SNR and has a resolution that approximates the original in-plane resolution in all directions. Experimental results on both simulated and real data sets are used to demonstrate the advantage of the proposed method.