Detección, seguimiento y análisis de personas en streams de vídeo

[ES] El presente TFG consiste en una aplicación para la detección de personas de cuerpo entero. La idea es aplicar este detector a las continuas imágenes recogidas en tiempo real a través de una web-cam, o de un archivo con formato de vídeo que se encuentre ubicado en el propio sistema. El código está escrito en C++. Para conseguir este objetivo nos basamos en el uso conjunto de dos sistemas de detección ya existentes: primero, OpenCV, mediante un método de histograma de gradientes orientados, el cual ya proporciona propiamente un detector de personas que será aplicado a cada una de las imágenes del stream de vídeo; por otro lado, el detector facial de la librería Encara que se aplica a cada una de las detecciones de supuestas personas obtenidas en el método de OpenCV, para comprobar si hay una cara en la supuesta persona detectada. En caso de ser así, y de haber una cara más o menos correctamente situada, determinamos que es realmente una persona. Para cada persona detectada se guardan sus datos de situación en la imagen, en una lista, para posteriormente compararlos con los datos obtenidos en frames anteriores, e intentar hacer un  seguimiento de todas las personas. Visualmente se observaría como se va recuadrando cada persona con un color determinado aleatorio asignado a cada una, mientras se visualiza el vídeo. También se registra la hora y frame de aparición, y la hora y frame de salida, de cada persona detectada, quedando estos datos guardados tanto en un fichero de log, como en una base de datos. Los resultados son, bastante satisfactorios, aunque con posibilidades de mejora, ya que es un trabajo que permite combinar otras técnicas diferentes a las descritas. Debido a la complejidad de los métodos empleados se destaca la necesidad de alta capacidad de computación para poder ejecutar la aplicación en tiempo real sin ralentizaciones. 

Learning with Kernels on Graphs: DAG-based kernels, data streams and RNA function prediction.

In many application domains data can be naturally represented as graphs. When the application of analytical solutions for a given problem is unfeasible, machine learning techniques could be a viable way to solve the problem. Classical machine learning techniques are defined for data represented in a vectorial form. Recently some of them have been extended to deal directly with structured data. Among those techniques, kernel methods have shown promising results both from the computational complexity and the predictive performance point of view. Kernel methods allow to avoid an explicit mapping in a vectorial form relying on kernel functions, which informally are functions calculating a similarity measure between two entities. However, the definition of good kernels for graphs is a challenging problem because of the difficulty to find a good tradeoff between computational complexity and expressiveness. Another problem we face is learning on data streams, where a potentially unbounded sequence of data is generated by some sources. There are three main contributions in this thesis. The first contribution is the definition of a new family of kernels for graphs based on Directed Acyclic Graphs (DAGs). We analyzed two kernels from this family, achieving state-of-the-art results from both the computational and the classification point of view on real-world datasets. The second contribution consists in making the application of learning algorithms for streams of graphs feasible. Moreover,we defined a principled way for the memory management. The third contribution is the application of machine learning techniques for structured data to non-coding RNA function prediction. In this setting, the secondary structure is thought to carry relevant information. However, existing methods considering the secondary structure have prohibitively high computational complexity. We propose to apply kernel methods on this domain, obtaining state-of-the-art results.

Investing in Community Colleges of the Commonwealth: A Review of Funding Streams

Community colleges play a pivotal role in Massachusetts’ higher education system. They award associate’s degrees, offer an affordable pathway to a bachelor's degree, and provide access to higher education for underserved populations. Community colleges also offer workforce development and industry certification for those entering the world of work, those who are currently employed but need more training or retraining, and those who have been dislocated from their careers by the global economic crisis. To increase access to both college credits and improved work skills, community colleges provide remedial or developmental education to ensure all students can pursue a program of study (Dougherty & Townsend, 2006; Mullin, 2010b).

Biofilm responses to salmon carcass analog addition in central Idaho streams

Pacific salmon populations have declined due to human activity in the Pacific Northwest, resulting in decreased delivery of marine-derived nutrients to streams. Managers use artificial nutrient additions to increase juvenile salmon growth and survival and assume that added nutrients stimulate biofilm production, which propagates up the food web to juvenile salmon. We assessed biofilm responses (standing crop, nutrient limitation, and metabolism) to experimental additions of salmon carcass analog in tributaries of the Salmon River, Idaho in 2010 and 2011. Biofilm standing crop and nutrient limitation did not respond to analog, but primary productivity and respiration increased in the subset of streams where they were measured. Discrepancies between biofilm productivity and standing crop may occur if standing crop is constrained by physical and biological factors. Thus, conclusions about biofilm response to analog should not be based on standing crop alone and mitigation research may benefit from nutrient budgets of entire watersheds.

An Ecological and Chemical Assessment of Copper Contaminated Streams in the Keweenaw Peninsula

The copper mining boom in Michigan's Upper Peninsula ended in the mid-1960s, but the historical mining still affects the region to this day. Earlier studies conducted in the Keweenaw have shown that trace metals in the sediments negatively affect benthic macroinvertebrate populations. However, because the concentrations of trace metals that are observed to be toxic often differ significantly between the laboratory and the environment, a better method for determining toxic levels of trace metals in the natural environment is desirable in order to establish surface water quality guidelines that effectively protect aquatic life. There were four research objectives for this research project. First, to determine if trace-level concentrations of copper can result in detectable ecological impacts even in the presence of high dissolved organic carbon (DOC). Second, to determine if there is a "safe" concentration of total dissolved copper below which there is little to no ecological impairment. Third, to establish which streams in the Keweenaw Peninsula have been most impacted by elevated levels of total dissolved copper. Fourth, to use this information to evaluate revisions to the water quality criterion for copper that were recently proposed by the Michigan Department of Environmental Quality (MDEQ). In order to collect water quality and macroinvertebrate data, two sampling surveys of approximately 50 streams were completed in the spring and summer of 2012. Our findings demonstrate that negative ecological impacts can be detected even in the presence of high concentrations of DOC. The majority of surveyed streams showed evidence of total dissolved copper concentrations that were elevated above background levels. Our findings suggest that there are detectable negative impacts below the current water quality standard for copper in many Keweenaw streams. The diversity of benthic macroinvertebrates and the number of species present has been reduced as a result of exposure to copper. Additionally, the multimetric approach used by MDEQ is unable to detect copper impairment in local streams due to the use of several insensitive metrics. The proposed changes to the copper criterion would increase the amount of total dissolved copper allowable despite the fact that approximately 25% of streams sampled have aquatic chemistries that would leave them vulnerable to high levels of copper ions.

Investigating Research Streams of Conjoint Analysis: A Bibliometric Study

Conjoint analysis (CA) is one of the most important methods for preference elicitation. In this paper we investigate the intellectual structure within the conjoint analytical research community. Analyses based on single papers provide a method-based overview of streams of conjoint research. By using novel bibliometric techniques in this field we complement findings of existing reviews. We use co-citation and factor analysis of the most cited articles in SSCI to identify the most important articles and research streams. Seven research streams are revealed which are visualized by means of multidimensional scaling. Tables and graphics reveal the disciplinary affiliations of contributors to CA, the special structure within the classes as well as links between them.

Headward retreat of streams in the Late Oligocene to Early Miocene Swiss Alps

This study uses the widths, the spacing and the grain-size pattern of Oligo/Miocene alluvial fan conglomerates in the central segment of the Swiss Alpine foreland to reconstruct the topographic development of the Alps. These data are analysed with models of longitudinal stream profile development, to propose that the Alpine topography evolved from an early transient state where streams adjusted to rock uplift by headward retreat, to a mature phase where any changes in rock uplift were accommodated by vertical incision. The first stage comprises the time interval between ca 31 Ma and 22 Ma, when the Alpine streams deposited many small fans with a lateral spacing of <30 km in the north Alpine foreland. As the range evolved, the streams joined and the fans coalesced into a few large depositional systems with a lateral spacing of ca 80 to 100 km at 22 Ma. The models used here suggest that the overall elevation of the Alps increased rapidly within <5 Myr. The variability in pebble size increased either due to variations in sediment supply, enhanced orographic effects, or preferentially due to a change towards a stormier palaeoclimate. By 22 Ma, only two large rivers carried material into the foreland fans, suggesting that the major Alpine streams had established themselves. This second phase of stable drainage network was maintained until ca 5 Ma, when the uplift and erosion of the Molasse started and streams were redirected both in the Alps and in the foreland. This study illustrates that sedimentological archives of foreland basins can be used to reconstruct the chronology of the topographic development of mountain belts. It is suggested that the finite elevation of mountainous landscapes is reached early during orogeny and can be maintained for millions of years, provided that erosion is efficient.

West Antarctic Ice Streams

Solar heat is the acknowledged driving force for climatic change. However, ice sheets are also capable of causing climatic change. This property of ice sheets derives from the facts that ice and rock are crystalline whereas the oceans and atmosphere are fluids and that ice sheets are massive enough to depress the earth's crust well below sea level. These features allow time constants for glacial flow and isostatic compensation to be much larger than those for ocean and atmospheric circulation and therefore somewhat independent of the solar variations that control this circulation. This review examines the nature of dynamic processes in ice streams that give ice sheets their degree of independent behavior and emphasizes the consequences of viscoplastic instability inherent in anisotropic polycrystalline solids such as glacial ice. Viscoplastic instability and subglacial topography are responsible for the formation of ice streams near ice sheet margins grounded below sea level. As a result the West Antarctic marine ice sheet is inherently unstable and can be rapidly carved away by calving bays which migrate up surging ice streams. Analyses of tidal flexure along floating ice stream margins, stress and velocity fields in ice streams, and ice stream boundary conditions are presented and used to interpret ERTS 1 photomosaics for West Antarctica in terms of characteristic ice sheet crevasse patterns that can be used to monitor ice stream surges and to study calving bay dynamics.

Correction to Variations of Ice Bed Coupling Beneath and Beyond Ice Streams: The Force Balance

A geometrical force balance that links stresses to ice bed coupling along a flow band of an ice sheet was developed in 1988 for longitudinal tension in ice streams and published 4 years later. It remains a work in progress. Now gravitational forces balanced by forces producing tensile, compressive, basal shear, and side shear stresses are all linked to ice bed coupling by the floating fraction phi of ice that produces the concave surface of ice streams. These lead inexorably to a simple formula showing how phi varies along these flow bands where surface and bed topography are known: phi = h(O)/h(I) with h(O) being ice thickness h(I) at x = 0 for x horizontal and positive upslope from grounded ice margins. This captures the basic fact in glaciology: the height of ice depends on how strongly ice couples to the bed. It shows how far a high convex ice sheet (phi = 0) has gone in collapsing into a low flat ice shelf (phi = 1). Here phi captures ice bed coupling under an ice stream and h(O) captures ice bed coupling beyond ice streams.