Chye Hwang Yan

Fast segmentation of bone in CT images using 3D adaptive thresholding

Fast bone segmentation is often important in computer-aided medical systems. Thresholding-based techniques have been widely used to identify the object of interest (bone) against dark backgrounds. However, the darker areas that are often present in bone tissue may adversely affect the results obtained using existing thresholding-based segmentation methods. We propose an automatic, fast, robust and accurate method for the segmentation of bone using 3D adaptive thresholding. An initial segmentation is first performed to partition the image into bone and non-bone classes, followed by an iterative process of 3D correlation to update voxel classification. This iterative process significantly improves the thresholding performance. A post-processing step of 3D region growing is used to extract the required bone region. The proposed algorithm can achieve sub-voxel accuracy very rapidly. In our experiments, the segmentation of a CT image set required on average less than 10s per slice. This execution time can be further reduced by optimizing the iterative convergence process.

Computational biomechanical modelling of the lumbar spine using marching-cubes surface smoothened finite element voxel meshing

There is a need for the development of finite element (FE) models based on medical datasets, such as magnetic resonance imaging and computerized tomography in computation biomechanics. Direct conversion of graphic voxels to FE elements is a commonly used method for the generation of FE models. However, conventional voxel-based methods tend to produce models with jagged surfaces. This is a consequence of the inherent characteristics of voxel elements; such a model is unable to capture the geometries of anatomical structures satisfactorily. We have developed a robust technique for the automatic generation of voxel-based patient-specific FE models. Our approach features a novel tetrahedronization scheme that incorporates marching-cubes surface smoothing together with a smooth-distortion factor (SDF). The models conform to the actual geometries of anatomical structures of a lumbar spine segment (L3). The resultant finite element analysis (FEA) at the surfaces is more accurate compared to the use of conventional voxel-based generated FE models. In general, models produced by our method were superior compared to that obtained using the commercial software ScanFE.

Rapid surface registration of 3D volumes using a neural network approach

An automatic surface-based rigid registration system using a neural network representation is proposed. The system has been applied to register human bone structures for image-guided surgery. A multilayer perceptron neural network is used to construct a patient-specific surface model from pre-operative images. A surface representation function derived from the resultant neural network model is then employed for intra-operative registration. The optimal transformation parameters are obtained via an optimization process. This segmentation/registration system achieves sub-voxel accuracy comparable to that of conventional techniques, and is significantly faster. These advantages are demonstrated using image datasets of the calcaneus and vertebrae.

A component-oriented software toolkit for patient-specific finite element model generation

A component-oriented software system, i.BioMech (image-based biomechanical modeling) is proposed for generation of patient-specific finite element model. It applies a systematic software engineering approach to patient/subject-specific meshing and assignment of material properties. The prototype program is based on the component object model (COM), which enables ease of combination of existing mesh generation algorithms and material property assignment schemes, and incorporation of new ones. It also facilitates utilization by other programming languages or platforms. Data input comprises a series of medical images captured from the patient. The output is a patient-specific finite element model for computational analysis using commercially available finite element software. The prototype software system provides a framework to compare the different finite element mesh generation methods as well as schemes for material property assignment. Our focus is on patient/subject-specific modeling of the human vertebrae.

Accurate Measurement of Bone Mineral Density Using Clinical CT Imaging With Single Energy Beam Spectral Intensity Correction

Although dual-energy X-ray absorptiometry (DXA) offers an effective measurement of bone mineral density, it only provides a 2-D projected measurement of the bone mineral density. Clinical computed tomography (CT) imaging will have to be employed for measurement of 3-D bone mineral density. The typical dual energy process requires precise measurement of the beam spectral intensity at the 80 kVp and 120 kVp settings. However, this is not used clinically because of the extra radiation dosage and sophisticated hardware setup. We propose an accurate and fast approach to measure bone material properties with single energy scans. Beam hardening artifacts are eliminated by incorporating the polychromatic characteristics of the X-ray beam into the reconstruction process. Bone mineral measurement from single energy CT correction is compared with that of dual energy correction and the commonly used DXA. Experimental results show that single energy correction is compatible with dual energy CT correction in eliminating beam hardening artifacts and producing an accurate measurement of bone mineral density. We can then estimate Youngs modulus, yield stress, yield strain and ultimate tensile stress of the bone, which are important data for patient specific therapy planning.

Video segmentation: Propagation, validation and aggregation of a preceding graph

In this work, video segmentation is viewed as an efficient intra-frame grouping temporally reinforced by a strong inter-frame coherence. Traditional approaches simply regard pixel motions as another prior in the MRF-MAP framework. Since pixel pre-grouping is inefficiently performed on every frame, the strong correlation between inter-frame groupings is largely underutilized. We exploit the inter-frame correlation to propagate trustworthy groupings from the previous frame. A preceding graph is constructed and labeled for the previous frame. It is temporally propagated to the current frame and validated by similarity measures. All unlabeled subgraphs are spatially aggregated for the final grouping. Experimental results show that the proposed approach is highly efficient for spatio-temporal segmentation. It makes good use of temporal correlation and produces satisfactory grouping results.

A neural network approach for 3D surface modeling and registration

Surface based registration is commonly used in image aided surgery. This technique is extremely computationally expensive due to (1) the number of iterations required to search through the large parameter space and (2) the heavy computational load needed for determining the cost function (the distance between two surfaces). This is the main obstacle in pushing surface based registration for image guided surgery, where near real time registration is needed. Most attempts to reduce the computational burden, e.g., gradient descent and ICP, have been targeted at reducing the number of iterations for the optimization. In this paper, we propose to use a neural network to model the surface of the reference structure. This not only provides an accurate model for the surface but also a fast method for computing the cost function. For CT-CT spine registration, the time taken to register two spine surfaces is about 10 times faster compared to the commonly used triangular mesh modeling with similar registration accuracy.

Accurate and fully automatic 3D registration of spinal images using normalized mutual information

Automatic and accurate multi-modality (CT/MRI) image registration is an important part of image guided surgery, pre-surgery planning and post-surgery evaluation. Surface based registration is commonly used for registration of CT and MRI images of bone. Surface extraction from CT and MRI datasets is a pre-requisite for the registration. It is known that it is not possible to achieve fully automatic, accurate and complete segmentation of the spine from MRI dataset. Thus surface based registration for CT-MRI spine datasets cannot be fully automated. In this paper, we investigate the use of normalized mutual information as a method for fully automatic and accurate registration of CT-MRI spine datasets. We have compared the registration results with those from the surface based registration. Our results are promising and show that normalized mutual information can be used to implement fully automatic and accurate registration for CT and MRI images of the spine.

Biomechanical Modeling of Bone-Needle Interaction for Haptic Rendering in Needle Insertion Simulation

Medical simulators are increasingly being used for surgical training. For interactive surgical simulation involving haptic rendering, the force at the needle tip has to be computed very fast. We are developing biomechanical models for bone needle insertion. The cortical bone can be regarded as a dense form of cancellous bone that can be modeled using a linear elastic material. The porosity of the bone determines the resistance felt as the user inserts the needle into the bone. The bar element method that represents each trabecular bone as a FE beam is most computationally efficient. With 1000 FE elements, the computed force feedback were close to the insertion force measured during experiments. However, the extended bar element method may be the more appropriate choice for taking into consideration the trabecular distribution and hence, inhomogeneous of bone. The simulation studies on bone-needle interaction also showed that a diamond bevel needle may penetrate the bone with less force

Intensity-Based volumetric registration of contrast-enhanced MR breast images

In this paper, we propose a fast intensity-based registration algorithm for the analysis of contrast-enhanced breast MR images. Motion between pre-contrast and post-contrast images has been modeled by a combination of rigid transformation and free-form deformation. By modeling the conditional probability function to be Gaussian and considering the normalized mutual information (NMI) criterion, we create a pair of auxiliary images to speed up the registration process. The auxiliary images are registered to the actual images by optimizing the simple sum of squared difference (SSD) criterion. The overall registration is achieved by linearly combining the deformation observed in the auxiliary images. One well-known problem of non-rigid registration of contrast enhanced images is the contraction of enhanced lesion volume. We address this problem by rejecting the intensity outliers from registration. Results have shown that our method could achieve accurate registration of the data while successfully prevent the contraction of the contrast enhanced lesion volume.