Adaptive Fusion Of Particle Filtering And Spatio-Temporal Motion Energy For Human Tracking

Object tracking is an active research area nowadays due to its importance in human computer interface, teleconferencing and video surveillance. However, reliable tracking of objects in the presence of occlusions, pose and illumination changes is still a challenging topic. In this paper, we introduce a novel tracking approach that fuses two cues namely colour and spatio-temporal motion energy within a particle filter based framework. We conduct a measure of coherent motion over two image frames, which reveals the spatio-temporal dynamics of the target. At the same time, the importance of both colour and motion energy cues is determined in the stage of reliability evaluation. This determination helps maintain the performance of the tracking system against abrupt appearance changes. Experimental results demonstrate that the proposed method outperforms the other state of the art techniques in the used test datasets.

Bio-port Free Energy City

In an age of depleting oil reserves and increasing energy demand, humanity faces a stalemate between environmentalism and politics, where crude oil is traded at record highs yet the spotlight on being ‘green’ and sustainable is stronger than ever. A key theme on today’s political agenda is energy independence from foreign nations, and the United Kingdom is bracing itself for nuclear renaissance which is hoped will feed the rapacious centralised system that the UK is structured upon. But what if this centralised system was dissembled, and in its place stood dozens of cities which grow and monopolise from their own energy? Rather than one dominant network, would a series of autonomous city-based energy systems not offer a mutually profitable alternative? Bio-Port is a utopian vision of a ‘Free Energy City’ set in Liverpool, where the old dockyards, redundant space, and the Mersey Estuary have been transformed into bio-productive algae farms. Bio-Port Free Energy City is a utopian ideal, where energy is superfluous; in fact so abundant that meters are obsolete. The city functions as an energy generator and thrives from its own product with minimal impact upon the planet it inhabits. Algaculture is the fundamental energy source, where a matrix of algae reactors swamp the abandoned dockyards; which themselves have been further expanded and reclaimed from the River Mersey. Each year, the algae farm is capable of producing over 200 million gallons of bio-fuel, which in-turn can produce enough electricity to power almost 2 million homes. The metabolism of Free-Energy City is circular and holistic, where the waste products of one process are simply the inputs of a new one. Livestock farming – once traditionally a high-carbon countryside exercise has become urbanised. Cattle are located alongside the algae matrix, and waste gases emitted by farmyards and livestock are largely sequestered by algal blooms or anaerobically converted to natural gas. Bio-Port Free Energy City mitigates the imbalances between ecology and urbanity, and exemplifies an environment where nature and the human machine can function productively and in harmony with one another. According to James Lovelock, our population has grown in number to the point where our presence is perceptibly disabling the planet, but in order to reverse the effects of our humanist flaws, it is vital that new eco-urban utopias are realised.

Robust Multimodal Person Identification With Limited Training Data

This paper presents a novel method of audio-visual feature-level fusion for person identification where both the speech and facial modalities may be corrupted, and there is a lack of prior knowledge about the corruption. Furthermore, we assume there are limited amount of training data for each modality (e.g., a short training speech segment and a single training facial image for each person). A new multimodal feature representation and a modified cosine similarity are introduced to combine and compare bimodal features with limited training data, as well as vastly differing data rates and feature sizes. Optimal feature selection and multicondition training are used to reduce the mismatch between training and testing, thereby making the system robust to unknown bimodal corruption. Experiments have been carried out on a bimodal dataset created from the SPIDRE speaker recognition database and AR face recognition database with variable noise corruption of speech and occlusion in the face images. The system's speaker identification performance on the SPIDRE database, and facial identification performance on the AR database, is comparable with the literature. Combining both modalities using the new method of multimodal fusion leads to significantly improved accuracy over the unimodal systems, even when both modalities have been corrupted. The new method also shows improved identification accuracy compared with the bimodal systems based on multicondition model training or missing-feature decoding alone.

Decision Making by Credal Nets

Credal nets are probabilistic graphical models which extend Bayesian nets to cope with sets of distributions. This feature makes the model particularly suited for the implementation of classifiers and knowledge-based systems. When working with sets of (instead of single) probability distributions, the identification of the optimal option can be based on different criteria, some of them eventually leading to multiple choices. Yet, most of the inference algorithms for credal nets are designed to compute only the bounds of the posterior probabilities. This prevents some of the existing criteria from being used. To overcome this limitation, we present two simple transformations for credal nets which make it possible to compute decisions based on the maximality and E-admissibility criteria without any modification in the inference algorithms. We also prove that these decision problems have the same complexity of standard inference, being NP^PP-hard for general credal nets and NP-hard for polytrees.

On the Virtualization of CUDA Based GPU Remoting on ARM and X86 Machines in the GVirtuS Framework

The astonishing development of diverse and different hardware platforms is twofold: on one side, the challenge for the exascale performance for big data processing and management; on the other side, the mobile and embedded devices for data collection and human machine interaction. This drove to a highly hierarchical evolution of programming models. GVirtuS is the general virtualization system developed in 2009 and firstly introduced in 2010 enabling a completely transparent layer among GPUs and VMs. This paper shows the latest achievements and developments of GVirtuS, now supporting CUDA 6.5, memory management and scheduling. Thanks to the new and improved remoting capabilities, GVirtus now enables GPU sharing among physical and virtual machines based on x86 and ARM CPUs on local workstations,computing clusters and distributed cloud appliances.

A marine reservoir correction database and on-line interface

Calibration is essential for interpretation of radiocarbon dates, especially when the 14C dates are compared to historical or climatic records with a different chronological basis. 14C ages of samples from the marine environment, such as shells or fish bones, or samples with a marine component, such as human bone in coastal regions, require an additional consideration because of the reservoir age of the ocean. While the pre-industrial global mean reservoir correction, R(t), is about 400 years, local variations (?R) can be several hundred years or more. ?R compilations on a global scale have been undertaken previously (Stuiver et al. 1986; Stuiver and Braziunas 1993), but have not been updated recently. Here we describe an on-line reservoir correction database accessed via mapping software. Rather than publishing a static ?R compilation, new data will be incorporated when it becomes available. The on-line marine reservoir correction database can be accessed at the website http://www.calib.org/.

A machine vision approach to the grading of crushed aggregate

The grading of crushed aggregate is carried out usually by sieving. We describe a new image-based approach to the automatic grading of such materials. The operational problem addressed is where the camera is located directly over a conveyor belt. Our approach characterizes the information content of each image, taking into account relative variation in the pixel data, and resolution scale. In feature space, we find very good class separation using a multidimensional linear classifier. The innovation in this work includes (i) introducing an effective image-based approach into this application area, and (ii) our supervised classification using wavelet entropy-based features.

The Essential Role of Human Databases for Learning in and Validation of Affectively Competent Agents.

This output is an invited and refereed chapter in the second of the two book length outputs resulting from the EU HUMAINE grant and follow-on grants. The book is in the OUP Affective Science Series and is intended to provide a theoretically oriented state of the art model for those working in the area of affective computing. Each chapter provides a synthesis of a specific area and presents new data/findings/approaches developed by the author(s) which take the area further. This chapter is in the section on ‘Approaches to developing expression corpora and databases.’ The chapter provides a critical synthesis of the issues involved in databases for affective computing and introduces the SEMAINE SAL Database, developed as an integral part of the EU SEMAINE Project (The Sensitive Agent Project 2008-2011) which is an interdisciplinary project. The project aimed to develop a computer interface that would allow a human to interact with an artificial agent in an emotional manner.

Proteomic profiling of human retinal pigment epithelium exposed to an advanced glycation-modified substrate

Purpose The retinal pigment epithelium (RPE) and underlying Bruch’s membrane undergo significant modulation during ageing. Progressive, age-related modifications of lipids and proteins by advanced glycation end products (AGEs) at this cell–substrate interface have been implicated in RPE dysfunction and the progression to age-related macular degeneration (AMD). The pathogenic nature of these adducts in Bruch’s membrane and their influence on the overlying RPE remains unclear. This study aimed to identify alterations in RPE protein expression in cells exposed to AGE-modified basement membrane (AGE-BM), to determine how this “aged” substrate impacts RPE function and to map the localisation of identified proteins in ageing retina. Methods Confluent ARPE-19 monolayers were cultured on AGE-BM and native, non-modified BM (BM). Following 28-day incubation, the proteome was profiled using 2-dimensional gel electrophoresis (2D), densitometry and image analysis was employed to map proteins of interest that were identified by electrospray ionisation mass spectrometry (ESI MS/MS). Immunocytochemistry was employed to localise identified proteins in ARPE-19 monolayers cultured on unmodified and AGE-BM and to analyze aged human retina. Results Image analysis detected altered protein spot densities between treatment groups, and proteins of interest were identified by LC ESI MS/MS which included heat-shock proteins, cytoskeletal and metabolic regulators. Immunocytochemistry revealed deubiquitinating enzyme ubiquitin carboxyterminal hydrolase-1 (UCH-L1), which was upregulated in AGE-exposed RPE and was also localised to RPE in human retinal sections. Conclusions This study has demonstrated that AGE-modification of basement membrane alters the RPE proteome. Many proteins are changed in this ageing model, including UCHL-1, which could impact upon RPE degradative capacity. Accumulation of AGEs at Bruch”s membrane could play a significant role in age-related dysfunction of the RPE.

IQ-motif selectivity in human IQGAP2 and IQGAP3: binding of calmodulin and myosin essential light chain

The IQGAP [IQ-motif-containing GAP (GTPase-activating protein)] family members are eukaryotic proteins that act at the interface between cellular signalling and the cytoskeleton. As such they collect numerous inputs from a variety of signalling pathways. A key binding partner is the calcium-sensing protein CaM (calmodulin). This protein binds mainly through a series of IQ-motifs which are located towards the middle of the primary sequence of the IQGAPs. In some IQGAPs, these motifs also provide binding sites for CaM-like proteins such as myosin essential light chain and S100B. Using synthetic peptides and native gel electrophoresis, the binding properties of the IQ-motifs from human IQGAP2 and IQGAP3 have been mapped. The second and third IQ-motifs in IQGAP2 and all four of the IQ-motifs of IQGAP3 interacted with CaM in the presence of calcium ions. However, there were differences in the type of interaction: while some IQ-motifs were able to form complexes with CaM which were stable under the conditions of the experiment, others formed more transient interactions. The first IQ-motifs from IQGAP2 and IQGAP3 formed transient interactions with CaM in the absence of calcium and the first motif from IQGAP3 formed a transient interaction with the myosin essential light chain MIc1sa. None of these IQ-motifs interacted with S100B. Molecular modelling suggested that all of the IQ-motifs, except the first one from IQGAP2 formed alpha-helices in solution. These results extend our knowledge of the selectivity of IQ-motifs for CaM and related proteins.