No description available.

No description available.

The advancement of computer architecture and chip design has set the stage for the deep learning revolution by supplying enormous computational power. In general, deep learning is built upon neural networks that can be regarded as a universal approximator of any mathematical function. In contrast to model-based machine learning where the representative features are designed by human engineers, deep learning enables the automatic discovery of desirable feature representations based on a data-driven manner. In this thesis, the applications of deep learning to visual content processing and analysis are discussed. For visual content processing, two novel approaches, named LCVSR and RawVSR, are proposed to address the common issues in the filed of Video Super-Resolution (VSR). In LCVSR, a new mechanism based on local dynamic filters via Locally Connected (LC) layers is proposed to implicitly estimate and compensate motions. It avoids the errors caused by the inaccurate explicit estimation of flow maps. Moreover, a global refinement network is proposed to exploit non-local correlations and enhance the spatial consistency of super-resolved frames. In RawVSR, the superiority of camera raw data (where the primitive radiance information is recorded) is harnessed to benefit the reconstruction of High-Resolution (HR) frames. The developed network is in line with the real imaging pipeline, where the super-resolution process serves as a pre-processing unit of ISP. Moreover, a Successive Deep Inference (SDI) module is designed in accordance with the architectural principle suggested by a canonical decomposition result for Hidden Markov Model (HMM) inference, and a reconstruction module is built with elaborately designed Attention based Residual Dense Blocks (ARDBs). For visual content analysis, a new approach, named PSCC-Net, is proposed to detect and localize image manipulations. It consists of two paths: a top-down path that extracts the local and global features from an input image, and a bottom-up path that first distinguishes manipulated images from pristine ones via a detection head, and then localizes forged regions via a progressive mechanism, where manipulation masks are estimated from small scales to large ones, each serving as a prior of the next-scale estimation. Moreover, a Spatio-Channel Correlation Module (SCCM) is proposed to capture both spatial and channel-wise correlations among extracted features, enabling the network to cope with a wide range of manipulation attacks. Extensive experiments validate that the proposed methods in this thesis have achieved the SOTA results and partially addressed the existing issues in previous works. / Dissertation / Doctor of Philosophy (PhD)

Over the coming decades, both food consumption and agricultural water use are expected to increase in response to growing populations. In light of these concerns, there has been a growing awareness and appreciation of the objectives of agricultural sustainability, which has the broad aim of securing food and water resources, without adversely affecting the environment or disenfranchising future generations. To ensure that irrigated fields optimize their water use towards a more sustainable application while remaining compliant with any imposed restrictions on access to water supplies (i.e. through water licensing), it is necessary to understand and quantify the water consumption of crops at appropriate spatial and temporal scales. Evaporation (E), also commonly referred to as evapotranspiration (ET), is the physical process of water vapor transport from the surface into the atmosphere. Evaporation can be estimated via interpretive modeling approaches that combine meteorological, radiative, vegetation, and other related properties to estimate land surface fluxes at any given time. The research presented herein aims to investigate the evaporative response of agricultural croplands across a range of spatial and temporal scales, with a focus on high-resolution and field-scale estimation. In particular, we explore the utility of novel CubeSat imagery to produce the highest spatial resolution (3 m) crop water use estimates ever retrieved from space. These high-resolution results are expanded through time by retrieving a daily evaporation product, offering an enhanced capacity to provide new insights into precision agriculture. The effects and implications of higher spatiotemporal resolutions are explored and contrasted against governmental satellite missions that operate at lower resolutions. An exploratory study on the use of unmanned aerial vehicles (UAVs) is also performed, specifically in the context of their capacity to mount miniaturized thermal sensors: with the accuracy and limitations of these sensors for deriving evaporation-type products examined. The overarching goal of this research is to advance the utility of space-based estimates of evaporation for precision agricultural applications, and to provide new high-spatial and temporal agricultural insights that can be directed towards improving water management and address food security concerns in a more sustainable manner.

The tasks of entity recognition through ontological commitment, fact extraction and organization with respect to a target schema, and entity deduplication have all been examined in recent years, and systems exist that can perform each individual task. A framework combining all these tasks, however, is still needed to accomplish the goal of automatically extracting and organizing biographical facts about persons found in historical documents into disambiguated entity records. We introduce FROntIER (Fact Recognizer for Ontologies with Inference and Entity Resolution) as the framework to recognize and extract facts using an ontology and organize facts of interest through inferring implicit facts using inference rules, a target ontology, and entity resolution. We give two case studies of FROntIER's performance over a few select pages from The Ely Ancestry [BEV02] and Index to The Register of Marriages and Baptisms in the Parish of Kilbarchan, 1649-1772 [Gra12].

Digital Elevation Models (DEMs) are computer-generated representations of the earth's surface. These surfaces can be used to predicted Line-of-Sight (LOS) radio propagation. DEM resolution can affect the results of this prediction. This study examines the effect of DEM resolution on accuracy by comparing varied resolution terrain data for a portion of Blacksburg, Virginia using the prediction of ESRI's ArcView® viewshed algorithm. Results show that resolutions between one-meter and thirty-meters have little effect on the aggregate accuracy of the viewshed. / Master of Science

No / Nigeria is Africa's largest oil producing country. Oil generates enormous wealth but also extensive and devastating conflict in the country. High Stakes and Stakeholders critically explores the oil conflict in Nigeria, its evolution, dynamics and most significantly, the interplay and consequences of high stake politics for the reproduction and persistence of the conflict. It presents a conceptual anatomy of state-oil industry-society relations and demonstrates how the embedded material interests and accumulation patterns of different stakeholders underlie, shape and complicate both the oil conflict and security. In addition, the book provides key insights into comparable conflicts elsewhere in the global south, developing a logical framework for resolving the oil conflict in Nigeria and for reforming the security sector.

<p> The Lower Cretaceous (Upper Albian) Viking Formation of central Alberta contains numerous linear sandbodies and fewer large irregularly-shaped sandbodies. Most studies to date have focused on individual sandbodies, leaving their interrelationships largely unknown. Developing a high-resolution allostratigraphy for the Viking of central Alberta allows mapping of regional bounding discontinuities and the definition of distinct allomembers. Placement of the Viking hydrocarbon fields within this framework permits an understanding of the exact stratigraphic relationships of the various fields. </p> <p> Examination of 120 cores and numerous well log correlations suggests the existence of four regionally mappable bounding discontinuities (BDl-4) which separate five distinct allomembers (I-V). Sandbodies within these allomembers were deposited in a variety of sedimentologically distinct environments. These include 'regional Viking' offshore to shoreface sandstones, prograding highstand shoreface sediments, transgressive incised shoreface sediments, and forced regressive, onlapping shoreface 'tongues'. The series of linear trending hydrocarbon fields from Joffre to Chain are also stratigraphically distinct. The sandbodies exist at five separate stratigraphic horizons and therefore are not all part of the same incised shoreface deposits. </p> <p> The four regional bounding discontinuities are interpreted as transgressive surfaces of erosion formed by four separate drops and subsequent rises of relative sea level. These fluctuations were greater than 30m and each complete cycle occurred over roughly 375,000 years. </p> / Thesis / Master of Science (MSc)

<p> A technique has been developed for constructing low temperature germanium bolometers for use as detectors in the far infrared. Their performance has been evaluated both by measuring their responsivity and noise and comparing these with the theoretical values and also by using the bolometer in conjunction with a far infrared spectrophotometer to obtain spectra showing the theoretical resolution of 0.11 cm^-1 at 55 cm^-1 expected for this instrument.</p> / Thesis / Master of Science (MSc)

<p> An important issue in the study of conflict behavior concerns the manner in which one party involved in a conflict situation, can influence his protagonist to resolve the conflict in a way Which is mutually rewarding (i.e. cooperative). The present thesis is addressed to this issue. In order to study the relationship between preasymptotic interaction patterns and stable states of cooperation and conflict, the data from a. variety of two-person, mixed motive games are organized such that: a} criteria are established for defining stable states of cooperation and conflict; b) preasymptotic interaction patterns are clearly distinguished from these stable, asymptotic states; and c) the role of each dyad member is considered separately. </p> <p> Organizing the data in the manner outlined above allows the description of strategies used by real subjects Which lead to high levels or cooperation and conflict. Dyads who attain a high level or cooperation are found to use a cautious trust strategy. This strategy consists of two components, a cooperative signalling component and a firmness component. The cooperative signalling component is operationalized in terms of the difference in the proportion or cooperative choices between one dyad member and the other. The firmness component is operationalized in terms of the level of retaliation against an uncooperative action (D reciprocity). Dyads who attain a high level or conflict. are characterlzed by an inappropriate signalling compcinent, and/or the absence of a firmness component. Data are presented which indicate how the requirements for a successful cooperative signalling component vary as a function of other variables. </p> <p> The implications of focusing on stable outcomes in a conflict situation, and organizing the data in the manner developed here, are discussed in terms of notions current in the psychological literature on conflict behavior. Data from the present thesis are used to argue that strategic variables are relatively more important factors in determining stable outcomes in a particular conflict situation, than either predispositional or personality factors. </p> / Thesis / Doctor of Philosophy (PhD)