Huafeng Liu

Meshfree particle method

Many of the computer vision algorithms have been posed in various forms of differential equations, derived from minimization of specific energy functionals, and the finite element representation and computation have become the de facto numerical strategies for solving these problems. However, for cases where domain mappings between numerical iterations or image frames involve large geometrical shape changes, such as deformable models for object segmentation and nonrigid motion tracking, these strategies may exhibit considerable loss of accuracy when the mesh elements become extremely skewed or compressed. We present a new computational paradigm, the meshfree particle method, where the object representation and the numerical calculation are purely based on the nodal points and do not require the meshing of the analysis domain. This meshfree strategy can naturally handle large deformation and domain discontinuity issues and achieve desired numerical accuracy through adaptive node and polynomial shape function refinement. We discuss in detail the element-free Galerkin method, including the shape function construction using the moving least square approximation and the Galerkin weak form formulation, and we demonstrate its applications to deformable model based segmentation and mechanically motivated left ventricular motion analysis.

Physiological-Model-Constrained Noninvasive Reconstruction of Volumetric Myocardial Transmembrane Potentials

Personalized noninvasive imaging of subject-specific cardiac electrical activity can guide and improve preventive diagnosis and treatment of cardiac arrhythmia. Compared to body surface potential (BSP) recordings and electrophysiological information reconstructed on heart surfaces, volumetric myocardial transmembrane potential (TMP) dynamics is of greater clinical importance in exhibiting arrhythmic details and arrythmogenic substrates inside the myocardium. This paper presents a physiological-model-constrained statistical framework to reconstruct volumetric TMP dynamics inside the 3-D myocardium from noninvasive BSP recordings. General knowledge of volumetric TMP activity is incorporated through the modeling of cardiac electrophysiological system, and is used to constrain TMP reconstruction. This physiological system is reformulated into a stochastic state-space representation to take into account model and data uncertainties, and nonlinear data assimilation is developed to estimate volumetric myocardial TMP dynamics from personal BSP data. Robustness of the presented framework to practical model and data errors is evaluated. Comparison of epicardial potential reconstructions with classical regularization-based approaches is performed on computational phantom regarding right bundle branch blocks. Further, phantom experiments on intramural focal activities and an initial real-data study on postmyocardial infarction demonstrate the potential of the framework in reconstructing local arrhythmic details and identifying arrhythmogenic substrates inside the myocardium.

Nonlinear analysis of the BOLD signal

The linearized filtering approach to the hemodynamic system is limited in capturing the inherent nonlinearities of physiological systems. The nonlinear estimation method therefore should be thought of as a natural way to access the nonlinear data assimilation problem. In this paper, we present a nonlinear filtering algorithm which is computationally expensive compared to the existing linearization filtering algorithms, for hemodynamic data assimilation, to address the deficiencies inherent to linearization. Simultaneous estimation of the physiological states and the system parameters have been demonstrated in a simulated and real data. The method provides more reasonable inference about the parameters of models for hemodynamic data assimilation.

A new compact position-sensitive PMT for scintillation detectors

A new compact position-sensitive photomultiplier tube (PS-PMT), Hamamatsu R7600-C12, has been developed for scintillation detectors. The PS-PMT has 11 stages of metal channel dynodes [2] and 6(X)+6(Y) crossed plate anodes in a 25.7 mm square/spl times/20 mm high metal can package, where the photo-sensitive area is 22 mm square. The performance of the PS-PMT was evaluated in terms of applicability to radiation imaging systems. In comparison with the former type of PS-PMT (Hamamatsu R5900-C8), the new PS-PMT provides smaller light spread and better position response. Also, by removing the flange at the bottom of the PS-PMT, the ratio of the effective area to the outward area is increased. The spatial resolution capability was demonstrated by imaging a stratified LSO array having an element of 1.8 mm/spl times/1.8 mm/spl times/10 mm. Each crystal element is clearly identified on the image map with 662 keV uniform gamma-ray irradiation. This paper describes the characteristics of the new PS-PMT and the experimental results for a gamma-ray imaging detector.

Using MicroPET Imaging in Quantitative Verification of the Acupuncture Effect in Ischemia Stroke Treatment

Acupuncture has been indispensable in Chinese medicine. However, its function still remains elusive. This paper studies the effect of acupuncture in ischemia stroke treatment using the Sprague Dawley rat animal model. We induced focal cerebral ischemia in rats using the middle cerebral artery occlusion (MCAO) procedure. For each rat in the real acupuncture group (n = 63), the sham acupoint treatment group (n = 62), and the blank control group (n = 30), we acquired 3-D fluorodeoxyglucose-microPET images at baseline, after MCAO, and after treatment, respectively. Then, we measured the changes of the injury-volume in the right hemisphere of these rats. The measurements showed that real acupuncture slightly reduced the injury-volume, sham acupoint treatment increased the injury-volume, and blank control had no obvious effect in reducing the injury-volume. Statistical tests also confirmed that acupuncture was more effective than random stimulus in improving the metabolic recovery after stroke.

Maximum a Posteriori Strategy for the Simultaneous Motion and Material Property Estimation of the Heart

In addition to its technical merits as a challenging nonrigid motion and structural integrity analysis problem, quantitative estimation of cardiac regional functions and material characteristics has significant physiological and clinical value. We developed a stochastic finite-element framework for the simultaneous recovery of myocardial motion and material parameters from medical image sequences with an extended Kalman filter approach, and we have shown that this simultaneous estimation strategy achieves more accurate and robust results than separated motion and material estimation efforts. In this paper, we present a new computational strategy for the framework based upon the maximum a posteriori estimation principles, realized through the extended Kalman smoother, that produces a sequence of kinematics state and material parameter estimation of the entire myocardium, including the endocardial, epicardial, and midwall tissues. The system dynamics equations of the heart are constructed using a biomechanical model with stochastic parameters, and the tissue material and deformation parameters are jointly estimated from the periodic imaging data. Noise-corrupted synthetic image sequences with known kinematics and material parameters are used to assess the accuracy and robustness of the framework. Experiments with canine magnetic resonance tagging and phase-contrast image sequences have been conducted with very promising results, as validated through comparison to the histological staining of postmortem myocardium.

Meshfree implementation of individualized active cardiac dynamics.

The cardiac physiome model has been proven to be useful for cardiac simulation, and has been more recently utilized to medical image analysis. To perform individualized analysis, structural images are necessary to provide subject-specific cardiac geometries. Although finite element methods have been extensively used for the spatial discretization of the myocardium, their complicated meshing procedures and element-based interpolation functions often result in algorithms which are either easy to implement but numerically inaccurate, or accurate but labor-intensive. In consequence, we have adopted the meshfree platform which provides element-free approximations for computational cardiology. Complicated volume meshing procedures are excluded, and no re-meshing is needed for improving spatial accuracy when deformation occurs. Furthermore, the polynomial bases for spatial approximation are not limited by the element structure. As a result, the meshfree platform is more adaptive to different cardiac geometries and thus beneficial to individualized analysis. In this paper, the cardiac physiome model tailored for medical image analysis is presented with its detailed 3D implementation using the meshfree methods. Experiments were performed to compare the meshfree methods with the finite element methods, and simulations were done on a cubical object to investigate the local behaviors of the cardiac physiome model, and on a human heart geometry extracted from a magnetic resonance image to verify its physiological plausibility.

Changes in topological organization of functional PET brain network with normal aging.

Recent studies about brain network have suggested that normal aging is associated with alterations in coordinated patterns of the large-scale brain functional and structural systems. However, age-related changes in functional networks constructed via positron emission tomography (PET) data are still barely understood. Here, we constructed functional brain networks composed of regions in younger (mean age years) and older (mean age years) age groups with PET data. younger and older healthy individuals were separately selected for two age groups, from a physical examination database. Corresponding brain functional networks of the two groups were constructed by thresholding average cerebral glucose metabolism correlation matrices of regions and analysed using graph theoretical approaches. Although both groups showed normal small-world architecture in the PET networks, increased clustering and decreased efficiency were found in older subjects, implying a degeneration process that brain system shifts from a small-world network to regular one along with normal aging. Moreover, normal senescence was related to changed nodal centralities predominantly in association and paralimbic cortex regions, e.g. increasing in orbitofrontal cortex (middle) and decreasing in left hippocampus. Additionally, the older networks were about equally as robust to random failures as younger counterpart, but more vulnerable against targeted attacks. Finally, methods in the construction of the PET networks revealed reasonable robustness. Our findings enhanced the understanding about the topological principles of PET networks and changes related to normal aging.

Meshfree Representation and Computation: Applications to Cardiac Motion Analysis

For medical image analysis issues where the domain mappings between images involve large geometrical shape changes, such as the cases of nonrigid motion recovery and inter-object image registration, the finite element methods exhibit considerable loss of accuracy when the elements in the mesh become extremely skewed or compressed. Therefore, algorithmically difficult and computationally expensive remeshing procedures must be performed in order to alleviate the problem. We present a general representation and computation framework which is purely based on the sampling nodal points and does not require the construction of mesh structure of the analysis domain. This meshfree strategy can more naturally handle very large object deformation and domain discontinuity problems. Because of its intrinsic h-p adaptivity, the meshfree framework can achieve desired numerical accuracy through adaptive node and polynomial shape function refinement with minimum extra computational expense. We focus on one of the more robust meshfree efforts, the element free Galerkin method, through the moving least square approximation and the Galerkin weak form formulation, and demonstrate its relevancy to medical image analysis problems. Specifically, we show the results of applying this strategy to physically motivated multiframe motion analysis, using synthetic data for accuracy assessment and for comparison to finite element results, and using canine magnetic resonance tagging and phase contrast images for cardiac kinematics recovery.

Dynamic Dual-Tracer PET Reconstruction

Although of important medical implications, simultaneous dual-tracer positron emission tomography reconstruction remains a challenging problem, primarily because the photon measurements from dual tracers are overlapped. In this paper, we propose a simultaneous dynamic dual-tracer reconstruction of tissue activity maps based on guidance from tracer kinetics. The dual-tracer reconstruction problem is formulated in a state-space representation, where parallel compartment models serve as continuous-time system equation describing the tracer kinetic processes of dual tracers, and the imaging data is expressed as discrete sampling of the system states in measurement equation. The image reconstruction problem has therefore become a state estimation problem in a continuous-discrete hybrid paradigm, and H infinity filtering is adopted as the estimation strategy. As H infinity filtering makes no assumptions on the system and measurement statistics, robust reconstruction results can be obtained for the dual-tracer PET imaging system where the statistical properties of measurement data and system uncertainty are not available a priori, even when there are disturbances in the kinetic parameters. Experimental results on digital phantoms, Monte Carlo simulations and physical phantoms have demonstrated the superior performance.