A spatially constrained fuzzy hyper-prototype clustering algorithm

We present in this paper a fuzzy clustering algorithm which can handle spatially constraint problems often encountered in pattern recognition. The proposed method is based on the notions of hyperplanes, the fuzzy c-means, and spatial constraints. By adding a spatial regularizer into the fuzzy hyperplane-based objective function, the proposed method can take into account additionally important information of inherently spatial data. Experimental results have demonstrated that the proposed algorithm achieves superior results to some other popular fuzzy clustering models, and has potential for cluster analysis in spatial domain.     

Supervised restoration of degraded medical images using multiple-point geostatistics

Reducing noise in medical images has been an important issue of research and development for medical diagnosis, patient treatment, and validation of biomedical hypotheses. Noise inherently exists in medical and biological images due to the acquisition and transmission in any imaging devices. Being different from image enhancement, the purpose of image restoration is the process of removing noise from a degraded image in order to recover as much as possible its original version. This paper presents a statistically supervised approach for medical image restoration using the concept of multiple-point geostatistics. Experimental results have shown the effectiveness of the proposed technique which has potential as a new methodology for medical and biological image processing.     

Unsupervised active contours driven by density distance and local fitting energy with applications to medical image segmentation

This study presents an efficient variational region-based active contour model for segmenting images without priori knowledge about the object or background. In order to handle intensity inhomogeneities and noise, we propose to integrate into the region-based local intensity model a global density distance inspired by the Bhattacharyya flow. The local term based on local information of segmented image allows the model to deal with bias field artifact, which arises in data acquisition processes. The global term, which is based on the density distance between the probability distribution functions of image intensity inside and outside the active contour, provides information for accurate segmentation, keeps the curve from spilling, and addresses noise in the image. Intensive 2D and 3D experiments on many imaging modalities of medical fields such as computed tomography, magnetic resonance imaging, and ultrasound images demonstrate the effectiveness of the model when dealing with images with blurred object boundary, intensity inhomogeneities, and noise.     

A numerical hydrodynamic and thermal characterization of an inter-strata liquid cooling solution for 3D ICs

A novel integrated thermal management solution is proposed to alleviate hot spots in a contemporary 3D IC architecture. The solution employs a series of integrated microchannels, interconnected through each stratum by through silicon fluidic vias (TSFVs), and permits the transfer of heat, via a coolant, from hot to cold zones. This microfluidic system is driven by an integrated AC electrokinetic pump embedded in the channel walls. Recent advancements in electrokinetic micropump technology have allowed greater increases in fluid velocity (mm/s) while operating within the voltage constraints of a 3D IC. This paper presents a 2D simulation of an electrokinetic micropump operating at V_pp = 1.5 V in a 40 μm channel and examines its velocity profile for six frequencies in the range 100 ≤ ω ≤ 100 MHz. An optimum frequency of 100 kHz was established within this range and this was further examined with a constant heat flux of 186 W/cm2 imposed on the wall for an inlet fluid temperature of 40°C. Temperature profiles are presented at the channel-silicon interface and compared with theory.     

A Knowledge Discovery System with Support for Model Selection and Visualization

The process of knowledge discovery in databases consists of several steps that are iterative and interactive. In each application, to go through this process the user has to exploit different algorithms and their settings that usually yield multiple models. Model selection, that is, the selection of appropriate models or algorithms to achieve such models, requires meta-knowledge of algorithm/model and model performance metrics. Therefore, model selection is usually a difficult task for the user. We believe that simplifying the process of model selection for the user is crucial to the success of real-life knowledge discovery activities. As opposed to most related work that aims to automate model selection, in our view model selection is a semiautomatic process, requiring an effective collaboration between the user and the discovery system. For such a collaboration, our solution is to give the user the ability to try various alternatives and to compare competing models quantitatively by performance metrics, and qualitatively by effective visualization. This paper presents our research on model selection and visualization in the development of a knowledge discovery system called D2MS. The paper addresses the motivation of model selection in knowledge discovery and related work, gives an overview of D2MS, and describes its solution to model selection and visualization. It then presents the usefulness of D2MS model selection in two case studies of discovering medical knowledge in hospital data—on meningitis and stomach cancer—using three data mining methods of decision trees, conceptual clustering, and rule induction.     

Generalized Linear Quadratic Gaussian Techniques for the Wind Benchmark Problem

This paper presents a comprehensive study of the stage-3 benchmark problem for response control of wind-excited tall buildings based on the linear quadratic gaussian (LQG) approach, and on its generalization, the k-cost-cumulant control method. For control design, the original nodal building model is first transformed into balanced modal space. The Hankel singular values (HSVs) and the power spectral density of the wind disturbances are calculated; and, based on them, a reduced-order model is derived by keeping the first six low-frequency modes. A balanced LQG (BLQG) controller is then determined by adopting the HSVs as a basis to choose the design weights. The main results of the paper are that the BLQG control design is able to come within 5–10% of the performance of the sample LQG controller supplied with the benchmark, but with control actions on the order of one-third less than the sample LQG, and with stability improvement features of a substantial nature over the range of stiffness perturbations specified in the benchmark. Finally, if the low authority BLQG controller is regarded as a one-cost-cumulant design, then with the appropriate use of the second-cost cumulant, the latter methodology is able to demonstrate how higher-authority controllers can give certain improvements in performance, but at the expense of significant investment in control action.     

Detection of cross-channel anomalies

The data deluge has created a great challenge for data mining applications wherein the rare topics of interest are often buried in the flood of major headlines. We identify and formulate a novel problem: cross-channel anomaly detection from multiple data channels. Cross-channel anomalies are common among the individual channel anomalies and are often portent of significant events. Central to this new problem is a development of theoretical foundation and methodology. Using the spectral approach, we propose a two-stage detection method: anomaly detection at a single-channel level, followed by the detection of cross-channel anomalies from the amalgamation of single-channel anomalies. We also derive the extension of the proposed detection method to an online settings, which automatically adapts to changes in the data over time at low computational complexity using incremental algorithms. Our mathematical analysis shows that our method is likely to reduce the false alarm rate by establishing theoretical results on the reduction of an impurity index. We demonstrate our method in two applications: document understanding with multiple text corpora and detection of repeated anomalies in large-scale video surveillance. The experimental results consistently demonstrate the superior performance of our method compared with related state-of-art methods, including the one-class SVM and principal component pursuit. In addition, our framework can be deployed in a decentralized manner, lending itself for large-scale data stream analysis.