Time-varying extreme pattern with dynamic models

This paper is concerned with the analysis of time series data with time-varying extreme pattern. This is achieved via a model formulation that considers separately the central part and the tail of the distributions, using a two-component mixture model. Extremes beyond a threshold are assumed to follow a generalized Pareto distribution (GPD). Temporal dependence is induced by allowing the GPD parameters to vary with time. Temporal variation and dependence is introduced at a latent level via the novel use of dynamic linear models (DLM). Novelty lies in the time variation of the shape and scale parameter of the resulting distribution. These changes in limiting regimes as time changes reflect better the data behavior, with important gains in estimation and interpretation. The central part follows a nonparametric mixture approach. The uncertainty about the threshold is explicitly considered. Posterior inference is performed through Markov Chain Monte Carlo (MCMC) methods. A variety of scenarios can be entertained and include the possibility of alternation of presence and absence of a finite upper limit of the distribution for different time periods. Simulations are carried out in order to analyze the performance of our proposed model. We also apply the proposed model to financial time series: returns of Petrobrás stocks and SP500 index. Results show advantage of our proposal over currently entertained models such as stochastic volatility, with improved estimation of high quantiles and extremes.     

iTunes University and the classroom: Can podcasts replace Professors?

iTunes University, a website with downloadable educational podcasts, can provide students the opportunity to obtain professors’ lectures when students are unable to attend class. To determine the effectiveness of audio lectures in higher education, undergraduate general psychology students participated in one of two conditions. In the lecture condition, participants listened to a 25-min lecture given in person by a professor using PowerPoint slides. Copies of the slides were given to aid note-taking. In the podcast condition, participants received a podcast of the same lecture along with the PowerPoint handouts. Participants in both conditions were instructed to keep a running log of study time and activities used in preparing for an exam. One week from the initial session students returned to take an exam on lecture content. Results indicated that students in the podcast condition who took notes while listening to the podcast scored significantly higher than the lecture condition. The impact of mobile learning on classroom performance is discussed.     

Locally Adapted Projections to Reduce Panorama Distortions

Displaying panoramic and wide angle views on a flat 2D display surface is necessarily prone to distortions. Perspective projections are limited to fairly narrow view angles. Cylindrical and spherical projections can show full 360° panoramas, but at the cost of curving straight lines, interfering with the perception of salient shapes in the scene.
In this paper, we introduce locally-adapted projections . Such projections are defined by a continuous projection surface consisting of both near-planar and curved parts. A simple and intuitive user interface allows the specification of regions of interest to be mapped to the near-planar parts, thereby reducing bending artifacts. We demonstrate the effectiveness of our approach on a variety of panoramic and wide angle images, including both indoor and outdoor scenes.     

The Kalman filter in the context of adaptive filter theory

Model‐based adaptive algorithms are usually derived with the help of the Wiener‐Hopf equation based on empirical statistics. They are often interpreted as an extension to their model‐independent counterparts, i.e. the stochastic‐gradient based adaptive filters. As a consequence, it is generally not considered worthwhile to show the analogy between Kalman filters and adaptive filters. This article pursues just these two goals. First, it tries to remove the notion that the Kalman filter is a complicated and unnecessary detour from the subject of adaptive filtering. Second, the advantage of a deeper insight into adaptive algorithms from Kalman's viewpoint emerges from our treatment. Based on a time‐varying FIR filter model, the Kalman filter is completely derived and serves as a general framework for the special case of model‐based adaptive filters. Copyright © 2004 John Wiley & Sons, Ltd.     

A closed-form solution to natural image matting

Interactive digital matting, the process of extracting a foreground object from an image based on limited user input, is an important task in image and video editing. From a computer vision perspective, this task is extremely challenging because it is massively ill-posed -- at each pixel we must estimate the foreground and the background colors, as well as the foreground opacity ("alpha matte") from a single color measurement. Current approaches either restrict the estimation to a small part of the image, estimating foreground and background colors based on nearby pixels where they are known, or perform iterative nonlinear estimation by alternating foreground and background color estimation with alpha estimation.In this paper we present a closed-form solution to natural image matting. We derive a cost function from local smoothness assumptions on foreground and background colors, and show that in the resulting expression it is possible to analytically eliminate the foreground and background colors to obtain a quadratic cost function in alpha. This allows us to find the globally optimal alpha matte by solving a sparse linear system of equations. Furthermore, the closed-form formula allows us to predict the properties of the solution by analyzing the eigenvectors of a sparse matrix, closely related to matrices used in spectral image segmentation algorithms. We show that high quality mattes for natural images may be obtained from a small amount of user input.     

The Shadow Meets the Mask: Pyramid-Based Shadow Removal

In this paper we propose a novel method for detecting and removing shadows from a single image thereby obtaining a high‐quality shadow‐free image. With minimal user assistance, we first identify shadowed and lit areas on the same surface in the scene using an illumination‐invariant distance measure. These areas are used to estimate the parameters of an affine shadow formation model. A novel pyramid‐based restoration process is then applied to produce a shadow‐free image, while avoiding loss of texture contrast and introduction of noise. Unlike previous approaches, we account for varying shadow intensity inside the shadowed region by processing it from the interior towards the boundaries. Finally, to ensure a seamless transition between the original and the recovered regions we apply image inpainting along a thin border. We demonstrate that our approach produces results that are in most cases superior in quality to those of previous shadow removal methods. We also show that it is possible to easily composite the extracted shadow onto a new background or modify its size and direction in the original image.     

Inter-comparison of wind profiles in the tropical boundary layer remotely sensed from GPS radiosonde and Doppler wind lidar

Winds play a very important role in the dynamics of the lower atmosphere, and there is a need to obtain vertical distribution of winds at high spatio-temporal resolution for various observational and modelling applications. Profiles of wind speed and direction obtained at two tropical Indian stations using a Doppler wind lidar during the Indian southwest monsoon season were inter-compared with those obtained simultaneously from GPS upper-air sounding (radiosonde). Mean wind speeds at Mahbubnagar (16.73° N, 77.98° E, 445 m above mean sea level) compare well in magnitude for the entire height range from 100 m to 2000 m. The mean difference in wind speed between the two techniques ranged from ?0.81 m s^?1 to +0.41 m s^?1, and the standard deviation of wind speed differences ranged between 1.03 m s^?1 and 1.95 m s^?1. Wind direction by both techniques compared well up to about 1200 m height and then deviated slightly from each other at heights above, with a standard deviation in difference of 19°–48°. At Pune (18^○32′ N, 73^○51′ E, 559 m above mean sea level), wind speed by both techniques matched well throughout the altitude range, but with a constant difference of about 1 m s^?1. The root mean square deviation in wind speed ranged from 1.0 to 1.6 m s^?1 and that in wind direction from 20° to 45°. The bias and spread in both wind speed and direction for the two stations were computed and are discussed. The study shows that the inter-comparison of wind profiles obtained by the two independent techniques is very good under conditions of low wind speeds, and they show larger deviation when wind speeds are large, probably due the drift of the radiosonde balloon away from the location.     

Ground-penetrating radar measurement of crop and surface water content dynamics

Ground-penetrating radar (GPR) with a suspended 1 GHz horn antenna was deployed for measurement of soil water contents and crop canopy properties over bare and electrically terminating surfaces. Surface reflection (SR) and signal propagation times (PT) were used to independently determine dielectric permittivity and water content of soil and canopy. Measured surface reflection coefficients progressively decreased with increasing canopy biomass according to Beer-Lambert type relationships. In contrast, PT measurements remained unaffected by canopy, and hence provided an accurate account of soil water content dynamics. Immediately after canopy removal, SR-based soil water content measurements were in close agreement with PT values. Canopy dielectric properties were inferred from canopy water contents (?_c-CWC) and canopy propagation times (?_c-CPT). Distinct canopy reflections were correlated with key canopy biophysical parameters. The study demonstrates the usefulness of a horn antenna GPR for characterization of vegetation canopy scattering, and for subcanopy water content measurements within a well-defined footprint, thereby offering a potential for calibration and verification of radar data collected from air- and spaceborne platforms.     

Apple pomace as food fortification ingredient: A systematic review and meta-analysis

The present review aimed to investigate and analyze the use of byproduct apple pomace as a fortification ingredient in different types of foods. The data obtained from English published articles found on Web of Science, Scopus, and Google Scholar in the period from 2007 to 2019 were used for making the table overview and meta-analysis of results described in those studies. The systematic review confirmed the importance of apple pomace use in the food industry due to the beneficial nutritional profile and ecological issue (waste management). The main attributes of apple pomace are high content of antioxidant compounds and dietary fibers. Dietary fibers from apple pomace significantly increased total fiber content in enriched products—meaning that the transfer of the fortification can be declared health beneficial. The conducted meta-analysis showed unambiguously the different influence of apple pomace addition according to fortified food commodity. The fortification drawbacks were noticeable in plant food products because darker and brownish color was not evaluated positively by panelists. Oppositely, color, as one of the main sensory characteristics, was beneficially affected in animal origin food. The sensory properties, including color, play an important role in product acceptance by consumers. Besides color, animal origin products fortified by apple pomace showed the most acceptable textural properties and oxidative stability.