Control and component selection for incremental motion systems

The main problem in the design of an incremental motion system is the heating of the motor which results from the start/stop type of operation. The question answered in this paper is how to select the components for this system and how to use them in order to minimize the operating temperature of the motor. In the analysis of the control selection, the optimal control was determined which minimizes the energy dissipation. It is shown how a triangular velocity profile, obtained using Bang-Bang control, corresponds to an efficiency of 0.75, whereas a slight change to trapezoidal velocity, increases the efficiency to 0.89 when compared with the optimal control. The optimal gear ratio will differ from the “inertial match” due to the presence of constant load torque. For a capstan drive, the inertia of the capstan must be much smaller than that of the motor, and not “inertially matched”.     

Arachne: Planning the telephone network at NYNEX

Arachne provides an excellent example of the success that can be derived from an effective integration of expert systems and optimization techniques. Automating IOF planning at NYNEX was an extremely challenging problem involving conflicting constraints: (i) the IOF network planning problem is quite complex, and (ii) the size of NYNEX s IOF network is quite large. In Arachne, we resolve these conflicting constraints by decomposing the planning task in such a way that only two types of subtask remain: (i) subtasks in which the size of the data is large, and the equipment cost of decisions small, an (ii) subtasks in which the data size is small and the equipment cost of decisions high. We apply heuristics to the former types of subtasks and optimization techniques to the latter. Combining heuristics and optimization techniques in this way provides a very good solution that optimizes the IOF network where capital investment is greatest. Copyright     

Blind image deconvolution using machine learning for three-dimensional microscopy

In this work, we propose a novel method for the regularization of blind deconvolution algorithms. The proposed method employs example-based machine learning techniques for modeling the space of point spread functions. During an iterative blind deconvolution process, a prior term attracts the point spread function estimates to the learned point spread function space. We demonstrate the usage of this regularizer within a Bayesian blind deconvolution framework and also integrate into the latter a method for noise reduction, thus creating a complete blind deconvolution method. The application of the proposed algorithm is demonstrated on synthetic and real-world three-dimensional images acquired by a wide-field fluorescence microscope, where the need for blind deconvolution algorithms is indispensable, yielding excellent results.     

Interactive Learning for Point‐Cloud Motion Segmentation

Segmenting a moving foreground (fg) from its background (bg) is a fundamental step in many Machine Vision and Computer Graphics applications. Nevertheless, hardly any attempts have been made to tackle this problem in dynamic 3D scanned scenes. Scanned dynamic scenes are typically challenging due to noise and large missing parts. Here, we present a novel approach for motion segmentation in dynamic point‐cloud scenes designed to cater to the unique properties of such data. Our key idea is to augment fg/bg classification with an active learning framework by refining the segmentation process in an adaptive manner. Our method initially classifies the scene points as either fg or bg in an un‐supervised manner. This, by training discriminative RBF‐SVM classifiers on automatically labeled, high‐certainty fg/bg points. Next, we adaptively detect unreliable classification regions (i.e. where fg/bg separation is uncertain), locally add more training examples to better capture the motion in these areas, and re‐train the classifiers to fine‐tune the segmentation. This not only improves segmentation accuracy, but also allows our method to perform in a coarse‐to‐fine manner, thereby efficiently process high‐density point‐clouds. Additionally, we present a unique interactive paradigm for enhancing this learning process, by using a manual editing tool. The user explicitly edits the RBF‐SVM decision borders in unreliable regions in order to refine and correct the classification. We provide extensive qualitative and quantitative experiments on both real (scanned) and synthetic dynamic scenes.     

Multi-physics MRI-based two-layer fluid–structure interaction anisotropic models of human right and left ventricles with different patch materials: Cardiac function assessment and mechanical stress analysis

Multi-physics right and left ventricle (RV/LV) fluid–structure interaction (FSI) models were introduced to perform mechanical stress analysis and evaluate the effect of patch materials on RV function. The FSI models included three different patch materials (Dacron scaffold, treated pericardium, and contracting myocardium), two-layer construction, fiber orientation, and active anisotropic material properties. The models were constructed based on cardiac magnetic resonance (CMR) images acquired from a patient with severe RV dilatation and solved by ADINA. Our results indicate that the patch model with contracting myocardium leads to decreased stress level in the patch area, improved RV function and patch area contractility.     

Orienting attention in visual working memory reduces interference from memory probes.

Given a changing visual environment, and the limited capacity of visual working memory (VWM), the contents of VWM must be in constant flux. Using a change detection task, the authors show that VWM is subject to obligatory updating in the face of new information. Change detection performance is enhanced when the item that may change is retrospectively cued 1 s after memory encoding and 0.5 s before testing. The retro-cue benefit cannot be explained by memory decay or by a reduction in interference from other items held in VWM. Rather, orienting attention to a single memory item makes VWM more resistant to interference from the test probe. The authors conclude that the content of VWM is volatile unless it receives focused attention, and that the standard change detection task underestimates VWM capacity. (PsycINFO Database Record (c) 2010 APA, all rights reserved)