Exploiting spatial-temporal context for trajectory based action video retrieval

Multimedia Tools and Applications - Retrieving videos with similar actions is an important task with many applications. Yet it is very challenging due to large variations across different videos....     

某车载式可升降微波塔的抗风分析

本文以某车载式可升降微波塔架为工程背景,分别运用拟静力法和计算流体动力学方法(CFD)计算该微波塔架的风荷载,利用SAP2000工程软件建立了该微波塔架的三维空间有限元模型,分析了静力风荷载作用下结构的位移响应和部分构件的内力响应,并对该种车载式结构提出了合理的改进意见。     

Nosocomial infection following cardiovascular surgery: comparison of two periods, 1987 vs. 1992

BACKGROUND: QT dispersion has been proposed as a simple, noninvasive measure for identifying patients at risk of postinfarction arrhythmia. It is assumed to reflect nonuniform ventricular repolarization, which, in turn, may result from regional differences in repolarization time as well as from localized activation delay. The aim of this study was to examine the relation between QT dispersion and intraventricular conduction abnormalities in patients with acute anterior wall myocardial infarction. METHODS AND RESULTS: Standard 12-lead electrocardiographic and 12-lead signal-averaged electrocardiographic recordings were performed in 25 patients with a first Q-wave anterior wall myocardial infarction. Measures calculated by using the 6 precordial (V1 through V6) leads for QT dispersion were (1) difference between maximum and minimum QT and QTc intervals and (2) standard deviation of QT and QTc intervals. Measures calculated from the signal-averaged electrocardiogram were (1) maximum filtered QRS duration; (2) mean; and (3) standard deviation of filtered QRS duration. No relation was found between any measure of filtered QRS duration and that of QT dispersion by using linear correlation analysis. Similarly, no significant association was demonstrated between the filtered QRS duration and corresponding QT interval measurements (total 131 leads). CONCLUSIONS: The lack of correlation between signal-averaged electrocardiogram indexes of slow intraventricular conduction and electrocardiogram variables of QT dispersion suggests an independent predictive value for the 2 methods in identifying patients at risk of postinfarction arrhythmia. This suggestion is further supported by the finding that altered activation sequence is an unlikely mechanism of QT dispersion in patients with acute myocardial infarction, as indicated by the lack of association between the filtered QRS duration and corresponding QT interval measurements.     

New insights into ice growth and melting modifications by antifreeze proteins

Antifreeze proteins (AFPs) evolved in many organisms, allowing them to survive in cold climates by controlling ice crystal growth. The specific interactions of AFPs with ice determine their potential applications in agriculture, food preservation and medicine. AFPs control the shapes of ice crystals in a manner characteristic of the particular AFP type. Moderately active AFPs cause the formation of elongated bipyramidal crystals, often with seemingly defined facets, while hyperactive AFPs produce more varied crystal shapes. These different morphologies are generally considered to be growth shapes. In a series of bright light and fluorescent microscopy observations of ice crystals in solutions containing different AFPs, we show that crystal shaping also occurs during melting. In particular, the characteristic ice shapes observed in solutions of most hyperactive AFPs are formed during melting. We relate these findings to the affinities of the hyperactive AFPs for the basal plane of ice. Our results demonstrate the relation between basal plane affinity and hyperactivity and show a clear difference in the ice-shaping mechanisms of most moderate and hyperactive AFPs. This study provides key aspects associated with the identification of hyperactive AFPs.     

Impurity Effects on Energy Gap in Bi<Subscript>2</Subscript>Sr<Subscript>2</Subscript>CaCu<Subscript>2</Subscript>O<Subscript>8+<Emphasis Type="Italic">y</Emphasis></Subscript> Investigated by Electronic Raman Scattering

The Raman scattering experiments were carried out in Zn-doped and Zn-free Bi₂Sr₂CaCu₂O_8+y with optimal hole concentration below and above T _c . The energy of pair-breaking peak in the B _1g Raman spectrum, corresponding to the magnitude of superconducting gap 2Δ ₀, is suppressed by 1% Zn-doping. In the normal state, the B _1g Raman spectrum for Zn-doped sample shows no pseudogap behaviour, suggesting that the pseudogap is strongly smeared by Zn-doping.     

Surface-induced dissociation of peptides and protein complexes in a quadrupole/time-of-flight mass spectrometer

A novel in-line surface-induced dissociation (SID) device was designed and implemented in a commercial QTOF instrument (Waters/Micromass QTOF II). This new setup allows efficient SID for a broad range of molecules. It also allows direct comparison with conventional collision-induced dissociation (CID) on the same instrument, taking advantage of the characteristics of QTOF instrumentation, including extended mass range, improved sensitivity, and better resolution compared with quadrupole analyzers and ion traps. Various peptides and a noncovalent protein complex have been electrosprayed and analyzed with the new SID setup. Here we present SID of leucine enkephalin, fibrinopeptide A, melittin, insulin chain-B, and a noncovalent protein complex from wheat, heat shock protein 16.9. The SID spectra were also compared to CID spectra. With the SID setup installed, ion transmission proved to be efficient. SID fragmentation patterns of peptides are, in general, similar to CID, with differences in the relative intensities of some peaks such as immonium ions, backbone cleavage b- versus y-type ions, and y- versus y-NH3 ions, suggesting enhanced accessibility to high-energy/secondary fragmentation channels with SID. Furthermore, these results demonstrate that the in-line SID setup is a valid substitute for CID, with potential advantages for activation of singly/multiply charged peptides and larger species such as noncovalent protein complexes.     

Fuzzy local Gaussian mixture model for brain MR image segmentation

Accurate brain tissue segmentation from magnetic resonance (MR) images is an essential step in quantitative brain image analysis. However, due to the existence of noise and intensity inhomogeneity in brain MR images, many segmentation algorithms suffer from limited accuracy. In this paper, we assume that the local image data within each voxel's neighborhood satisfy the Gaussian mixture model (GMM), and thus propose the fuzzy local GMM (FLGMM) algorithm for automated brain MR image segmentation. This algorithm estimates the segmentation result that maximizes the posterior probability by minimizing an objective energy function, in which a truncated Gaussian kernel function is used to impose the spatial constraint and fuzzy memberships are employed to balance the contribution of each GMM. We compared our algorithm to state-of-the-art segmentation approaches in both synthetic and clinical data. Our results show that the proposed algorithm can largely overcome the difficulties raised by noise, low contrast, and bias field, and substantially improve the accuracy of brain MR image segmentation.     

Modeling Solar Drying Rate of Wastewater Sludge

Efficient solar drying requires that the drying rate is quantitatively known as a function of the environment and the control. To develop a drying-rate model for wastewater sludge, data were collected at a solar drying installation in Füssen, Germany. In this solar dryer, wet sludge is uniformly spread over a concrete floor under a greenhouse-like transparent cover. The sludge is mixed mechanically several times a day by an autonomous robot (electric mole®), the structure is fan-ventilated horizontally, and the indoor air is mixed by electric fans. Data of evaporation rate, environmental conditions, and control operations were collected over three drying cycles. Evaporation rate via sludge sampling and via vapor balance across the structure compared favorably, justifying the use of hourly vapor-balance data. Four types of prediction models were considered: physical, additive, multiplicative, and neural network. The multiplicative model has been selected for potential implementation. The most important predictors of evaporation rate, for the conditions under consideration, were (1) solar radiation, (2) outdoor temperature, (3) ventilation rate, and (4) dry solids content of the sludge. Air mixing is an order of magnitude less effective (per unit of air discharge) than ventilation.     

Optimal Control of Solar Sludge Dryers

Optimal control (OC) methodology is used to develop a control policy for a batch-operated solar sludge dryer with ventilation as the dominant control. Batch performance criteria and instantaneous optimization criteria are developed for two economic environments: quota limited and capital limited. The general formulation, as well as simplifications resulting from the multiplicative form of the evaporation (drying) rate model, are presented. The multiplicative form results in a single constant number, called here “control intensity,” which is used to guide the on-line control decisions. These decisions turn out to be independent of the dry solids content (DSC) of the sludge. Further simplification, by assuming a strictly constant weather, is used to demonstrate the general effects of the economic and physical (weather) environment on the solution.

The OC approach is used to address, via simulation under realistic weather conditions, several design, operation, and pricing problems. In particular, the following have been considered: sizing of the ventilation fans, determination of a fair fee for sludge disposal, finding the best final DSC, and evaluating the effect of the price of electricity. Regarding the particular location represented by the data, it has been shown that the installed capacity of the ventilation fans should be increased and that there seems to be sufficient economic incentive for solar drying.     

Fast human action classification and VOI localization with enhanced sparse coding

Sparse coding which encodes the natural visual signal into a sparse space for visual codebook generation and feature quantization, has been successfully utilized for many image classification applications. However, it has been seldom explored for many video analysis tasks. In particular, the increased complexity in characterizing the visual patterns of diverse human actions with both the spatial and temporal variations imposes more challenges to the conventional sparse coding scheme. In this paper, we propose an enhanced sparse coding scheme through learning discriminative dictionary and optimizing the local pooling strategy. Localizing when and where a specific action happens in realistic videos is another challenging task. By utilizing the sparse coding based representations of human actions, this paper further presents a novel coarse-to-fine framework to localize the Volumes of Interest (VOIs) for the actions. Firstly, local visual features are transformed into the sparse signal domain through our enhanced sparse coding scheme. Secondly, in order to avoid exhaustive scan of entire videos for the VOI localization, we extend the Spatial Pyramid Matching into temporal domain, namely Spatial Temporal Pyramid Matching, to obtain the VOI candidates. Finally, a multi-level branch-and-bound approach is developed to refine the VOI candidates. The proposed framework is also able to avoid prohibitive computations in local similarity matching (e.g., nearest neighbors voting). Experimental results on both two popular benchmark datasets (KTH and YouTube UCF) and the widely used localization dataset (MSR) demonstrate that our approach reduces computational cost significantly while maintaining comparable classification accuracy to that of the state-of-the-art methods.