Jens Gregor

Computational Analysis and Improvement of SIRT

Iterative X-ray computed tomography (CT) algorithms have the potential for producing high-quality images but are computationally very demanding, especially when applied to high-resolution problems. Focusing on simultaneous iterative reconstruction technique (SIRT), we provide an eigenvalue based scheme for automatically determining a near-optimal value of the relaxation parameter. This accelerates the convergence rate of SIRT to the point where only half the number of iterations normally required is needed. We also modify the way SIRT uses preconditioning to solve a weighted least squares problem. The resulting algorithm, which we call PSIRT, is associated with a smaller memory footprint and calls for less data to be communicated in a distributed-memory implementation. Experimental residual norm and timing results are provided based on cone-beam micro-CT mouse data, including for an ordered subsets study.

Dynamic programming alignment of sequences representing cyclic patterns

String alignment by dynamic programming is generalized to include cyclic shift and corresponding optimal alignment cost for strings representing cyclic patterns. A guided search algorithm uses bounds on alignment costs to find all optimal cyclic shifts. The bounds are derived from submatrices of an initial dynamic programming matrix. Algorithmic complexity is analyzed for major stages in the search. The applicability of the method is illustrated with satellite DNA sequences and circularly permuted protein sequences. >

Indoor scene reconstruction from sets of noisy range images

The paper describes a system for indoor scene modeling from sets of noisy laser range images. We address several important aspects thereof: preprocessing, which includes image segmentation and planar model fitting; view registration, which is the method of determining the rigid transformation that describes the relative pose of the camera platform between views; and reconstruction, which is the subsequent integration or fusion of separate range images into a single 3D model. We give an empirical analysis, which demonstrates the efficacy of our plane-based registration strategy and results in real data that show the performance of the reconstruction system in its entirety.

Quantitative high-resolution microradiographic imaging of amyloid deposits in a novel murine model of AA amyloidosis

The mouse model of experimentally induced systemic AA amyloidosis is long established, well validated, and closely analogous to the human form of this disease. However, the induction of amyloid by experimental inflammation is unpredictable, inconsistent, and difficult to modulate. We have previously shown that murine AA amyloid deposits can be imaged using iodine-123 labeled SAP scintigraphy and report here substantial refinements in both the imaging technology and the mouse model itself. In this regard, we have generated a novel prototype of AA amyloid in which mice expressing the human interleukin 6 gene, when given amyloid enhancing factor, develop extensive and progressive systemic AA deposition without an inflammatory stimulus, i.e., a transgenic rapidly inducible amyloid disease (TRIAD) mouse. Additionally, we have constructed high-resolution micro single photon emission computed tomography (SPECT)/computed tomography (CT) instrumentation that provides images revealing the precise anatomic location of amyloid deposits labeled by radioiodinated serum amyloid P component (SAP). Based on reconstructed microSPECT/CT images, as well as autoradiographic, isotope biodistribution, and quantitative histochemical analyses, the 125I-labeled SAP tracer bound specifically to hepatic and splenic amyloid in the TRIAD animals. The ability to discern radiographically the extent of amyloid burden in the TRIAD model provides a unique opportunity to evaluate the therapeutic efficacy of pharmacologic compounds designed to inhibit fibril formation or effect amyloid resolution.

High-Resolution Neutron and X-Ray Imaging of Granular Materials

AbstractHigh spatial resolution (∼13.7 mm/pixel) neutron tomography was performed on partially water-saturated compacted silica sand specimens with two different grain morphologies (round and angular) at Helmholtz Zentrum Berlin using cold neutrons at the cold neutron radiography and tomography beam line. A specimen mixed with heavy water was imaged for contrast comparison purposes. Microfocus X-ray imaging was also performed on these specimens with slightly higher resolution (∼11.2 mm/pixel) using geometric magnification to locate the solid phase (silica particle boundaries) more precisely. Image processing was performed to remove unwanted gammas detected because of the gadox scintillator used for the high-resolution neutron imaging system. The visualization of solid, gas, and liquid phases for different grain morphologies is presented at the grain level. Using dual-modal contrast possible from simultaneous use of neutrons and X-rays, the authors introduce, for the first time, an improved ability to dist...

Micro-Imaging of Amyloid in Mice

Scintigraphic imaging of radioiodinated serum amyloid P-component is a proven method for the clinical detection of peripheral amyloid deposits (Hawkins et al., 1990). However, the inability to perform comparably high-resolution studies in experimental animal models of amyloid disease has impacted not only basic studies into the pathogenesis of amyloidosis but also in the preclinical in vivo evaluation of potential anti-amyloid therapeutic agents. We have developed microimaging technologies, implemented novel computational methods, and established protocols to generate high-resolution images of amyloid deposits in mice. (125)I-labeled serum amyloid P component (SAP) and an amyloid-fibril reactive murine monoclonal antibody (designated 11-1F4) have been used successfully to acquire high-resolution single photon emission computed tomographic (SPECT) images that, when fused with x-ray computed tomographic (CT) data, have provided precise anatomical localization of secondary (AA) and primary (AL) amyloid deposits in mouse models of these diseases. This chapter will provide detailed protocols for the radioiodination and purification of amyloidophilic proteins and the generation of mouse models of AA and AL amyloidosis. A brief description of the available hardware and the parameters used to acquire high-resolution microSPECT and CT images is presented, and the tools used to perform image reconstruction and visualization that permit the analysis and presentation of image data are discussed. Finally, we provide established methods for measuring organ- and tissue-specific activities with which to corroborate the microSPECT and CT images.

Fast Feldkamp reconstruction based on focus of attention and distributed computing

The Feldkamp algorithm is widely accepted as a practical conebeam reconstruction method for three‐dimensional x‐ray computed tomography. We introduce focus of attention, an effective and simple to implement datadriven preprocessing scheme, for identifying a convex subset of voxels that include all those relevant to the object under study. By concentrating on this subset of voxels during reconstruction, we reduce the computational demands of the Feldkamp algorithm correspondingly. To achieve further speed‐up, all computations are distributed across a cluster of inexpensive, dual‐processor PCs. We present experimental work based on mouse data obtained from the MicroCAT which is a high‐resolution x‐ray computed tomography system for small animal imaging. This work shows that focus of attention can cut the overall computation time in half without affecting the image quality. The method is general by nature and can easily be adapted to apply to other geometries and modalities as well as to iterative reconstruction algorithms.

A maximum-likelihood surface estimator for dense range data

Describes how to estimate 3D surface models from dense sets of noisy range data taken from different points of view, i.e., multiple range maps. The proposed method uses a sensor model to develop an expression for the likelihood of a 3D surface, conditional on a set of noisy range measurements. Optimizing this likelihood with respect to the model parameters provides an unbiased and efficient estimator. The proposed numerical algorithms make this estimation computationally practical for a wide variety of circumstances. The results from this method compare favorably with state-of-the-art approaches that rely on the closest-point or perpendicular distance metric, a convenient heuristic that produces biased solutions and fails completely when surfaces are not sufficiently smooth, as in the case of complex scenes or noisy range measurements. Empirical results on both simulated and real ladar data demonstrate the effectiveness of the proposed method for several different types of problems. Furthermore, the proposed method offers a general framework that can accommodate extensions to include surface priors, more sophisticated noise models, and other sensing modalities, such as sonar or synthetic aperture radar.

Efficient Dynamic Programming Alignment of Cyclic Strings by Shift Elimination

Abstract Optimal alignment of two strings of length m and n is computed in time O ( mn ) by dynamic programming. When the strings represent cyclic patterns, the alignment computation must consider all possible shifts and the computation complexity increases accordingly. We present an algorithm for efficient dynamic programming alignment of cyclic strings which uses a previously established channeling technique to reduce the complexity of each alignment and a new shift elimination technique to reduce the number of alignments carried out. The result is a data-dependent time complexity that varies between O (2 mn ) and O ( mn log 2 n ). An experimental evaluation is provided using randomly generated sequences.

Probabilistic suffix models for API sequence analysis of Windows XP applications

Given the pervasive nature of malicious mobile code (viruses, worms, etc.), developing statistical/structural models of code execution is of considerable importance. We investigate using probabilistic suffix trees (PSTs) and associated suffix automata (PSAs) to build models of benign application behavior with the goal of subsequently being able to detect malicious applications as anything that deviates therefrom. We describe these probabilistic suffix models and present new generic analysis and manipulation algorithms. The models and the algorithms are then used in the context of API (i.e., system call) sequences realized by Windows XP applications. The analysis algorithms, when applied to traces (i.e., sequences of API calls) of benign and malicious applications, aid in choosing an appropriate modeling strategy in terms of distance metrics and consequently provide classification measures in terms of sequence-to-model matching. We give experimental results based on classification of unobserved traces of benign and malicious applications against a suffix model trained solely from traces generated by a small set of benign applications.