Use of fuzzy-logic-inspired features to improve bacterialrecognition through classifier fusion

Escherichia coli O157:H7 has been found to cause serious health problems. Traditional methods to identify the organism are quite slow, pulsed-held gel electrophoresis (PFGE) images contain "banding pattern" information which can be used to recognize the bacteria. A fuzzy logic rule-based system is used as a guide to find a good feature set for the recognition of E. coli O157:H7. While the fuzzy rule-based system achieved good recognition, the human inspired features used in the rules were incorporated into a multiple neural network fusion approach which gave excellent separation of the target bacteria. The fuzzy integral was utilized in the fusion of neural networks trained with different feature sets to reach an almost perfect classification rate of E. coli O157:H7 PFGE patterns made available for the experiments.     

Mode-locked laser cavities with a single prism for dispersion compensation

We demonstrate the use of a single prism for adjustable dispersion compensation in a mode-locked laser cavity, instead of the standard approach with a prism pair. A simple model based on the prism-pair configuration is presented to determine the group-velocity dispersion by use of ray optics to trace the wavelength-dependent optical axes through the cavity. We experimentally demonstrated this concept with a passively mode-locked diode-pumped Nd:glass laser producing 200-fs pulses with a 200-mW average output power, using only one intracavity prism. The advantages of such a cavity design are simple alignment, reduced loss, and lossless wavelength tunability This technique can be generalized to other angularly dispersive elements such as prismatic output couplers.     

Capillary HPLC-NMR Coupling: High-Resolution (1)H NMR Spectroscopy in the Nanoliter Scale

Coupling HPLC and NMR is one of the most powerful techniques for simultaneous separation and structural elucidation of unknown compounds in mixtures. To date, however, minimizing the detection volume, as is required when coupling NMR with miniaturized separation techniques, has been accompanied by a dramatic loss in resolution of the NMR spectra. Here, we report on the coupling of gradient capillary HPLC with on-column, high-resolution NMR detection. On-line stopped-flow and static (1)H NMR spectra were acquired with capillary columns of 75-315 μm i.d. With detection over a length of 1.2 cm, cell volumes cover a range of 50-900 nL. An on-line-detected NMR separation of dansylated amino acids was carried out in a 315 μm i.d. fused silica capillary packed to a length of 12 cm with C(18) stationary phase. The low solvent consumption makes the use of fully deuterated solvents economically feasible. NMR spectra with resolution on the order of 3 Hz were obtained using a 50 nL detection cell to measure 1.1 nmol of dansylated γ-aminobutyric acid under static conditions in a 75 μm i.d. capillary.     

Thin-film designs by simulated annealing

With the increasing power of computers, new methods in synthesis of optical multilayer systems have appeared. Among these, the simulated-annealing algorithm has proved its efficiency in several fields of physics. We propose to show its performances in the field of optical multilayer systems through different filter designs.     

Intradermal and subcutaneous leiomyosarcoma: a clinicopathological and immunohistochemical study of 41 cases

Superficial leiomyosarcomas are rare tumours. The lesions confined to the dermis, contrary to those involving the subcutis, have been reported to carry a favourable prognosis. A retrospective study of 41 consecutive cases of surgically treated intradermal and subcutaneous leiomyosarcomas was undertaken in order to determine the prognostic factors that may influence the survival of these patients. Seven tumours were predominantly intradermal and 34 involved the subcutaneous tissue. Fifty-four percent of the tumours were located in the lower extremities. All cases stained positively for smooth muscle antigen and 66% for desmin. The tumours were classified with regard to tumour grade I (low grade, 3%), II (intermediate, 12%), IIIA (high grade, 46%) and IIIB (high grade, 39%). In all patients, follow-up information was available. Mean follow-up time was 5 years. The patients with intradermal tumours were all alive without signs of recurrence, whereas 14 of those with leiomyosarcomas involving the subcutis have died with pulmonary metastases. Our study confirms that "pure" intradermal leiomyosarcomas independent of tumour grade behave in a benign fashion, probably due to small tumour size. Tumour size > or = 5 cm, deep localization with fascia involvement, and high malignancy grade (IIIB) were found to deteriorate survival based on a univariate analysis. However, in a multivariate analysis only tumour size was found to be an independent prognostic factor.     

Socio-technical IS design science research: developing design theory for IS integration management

Design science research is an essential part of IS research since the field should not only try to understand how the world is, but also how to change it. We argue that the aim of IS design science research should be to develop practical knowledge not only for the design of novel information technology (IT), but also for IS governance and management. Whereas at least some methodological support exists for researchers engaged in IT-centric design science research, limited support is available for researchers who want to develop design knowledge and theory for IS governance and management. To overcome this shortcoming, we suggest a socio-technical IS design science research approach. The approach has four main activities: (1) identifying problem situations and desired outcomes, (2) reviewing extant theories, knowledge and data, (3) proposing/refining design theory and knowledge, and (4) testing design theory and knowledge. The applicability and usefulness of the proposed approach is shown by means of a design science research project concerning IS integration management in the context of mergers and acquisitions.     

Distinguishing attacks on stream ciphers based on arrays of pseudo-random words

In numerous modern stream ciphers, the internal state consists of a large array of pseudo-random words, while the output key-stream is a relatively simple function of the state. It has been heuristically shown in several situations [3], [8], [9], [10], [11] and [14] that this structure may lead to distinguishing attacks on the cipher. In this note we present a more rigorous treatment of this structural attack. First, we present a rigorous proof of the main probabilistic claim behind it in the basic cases. We then apply it concretely to the cipher sn3 [12], and demonstrate that the heuristic assumptions of the attack are remarkably precise in more complicated cases.     

Efficient high order matching

We present a computational approach to high-order matching of data sets in IR(d). Those are matchings based on data affinity measures that score the matching of more than two pairs of points at a time. High-order affinities are represented by tensors and the matching is then given by a rank-one approximation of the affinity tensor and a corresponding discretization. Our approach is rigorously justified by extending Zass and Shashua's hypergraph matching to high-order spectral matching. This paves the way for a computationally efficient dual-marginalization spectral matching scheme. We also show that, based on the spectral properties of random matrices, affinity tensors can be randomly sparsified while retaining the matching accuracy. Our contributions are experimentally validated by applying them to synthetic as well as real data sets.     

Textures on Rank‐1 Lattices

Storing textures on orthogonal tensor product lattices is predominant in computer graphics, although it is known that their sampling efficiency is not optimal. In two dimensions, the hexagonal lattice provides the maximum sampling efficiency. However, handling these lattices is difficult, because they are not able to tile an arbitrary rectangular region and have an irrational basis. By storing textures on rank‐1 lattices, we resolve both problems: Rank‐1 lattices can closely approximate hexagonal lattices, while all coordinates of the lattice points remain integer. At identical memory footprint texture quality is improved as compared to traditional orthogonal tensor product lattices due to the higher sampling efficiency. We introduce the basic theory of rank‐1 lattice textures and present an algorithmic framework which easily can be integrated into existing off‐line and real‐time rendering systems.