Face Recognition from Video Using the Generic Shape-Illumination Manifold

In spite of over two decades of intense research, illumination and pose invariance remain prohibitively challenging aspects of face recognition for most practical applications. The objective of this work is to recognize faces using video sequences both for training and recognition input, in a realistic, unconstrained setup in which lighting, pose and user motion pattern have a wide variability and face images are of low resolution. In particular there are three areas of novelty: (i) we show how a photometric model of image formation can be combined with a statistical model of generic face appearance variation, learnt offline, to generalize in the presence of extreme illumination changes; (ii) we use the smoothness of geodesically local appearance manifold structure and a robust same-identity likelihood to achieve invariance to unseen head poses; and (iii) we introduce an accurate video sequence “reillumination” algorithm to achieve robustness to face motion patterns in video. We describe a fully automatic recognition system based on the proposed method and an extensive evaluation on 171 individuals and over 1300 video sequences with extreme illumination, pose and head motion variation. On this challenging data set our system consistently demonstrated a nearly perfect recognition rate (over 99.7%), significantly outperforming state-of-the-art commercial software and methods from the literature

Multitasking: multiple, domain-specific cognitive functions in a virtual environment

Multitasking among three or more different tasks is a ubiquitous requirement of everyday cognition, yet rarely is it addressed in research on healthy adults who have had no specific training in multitasking skills. Participants completed a set of diverse subtasks within a simulated shopping mall and office environment, the Edinburgh Virtual Errands Test (EVET). The aim was to investigate how different cognitive functions, such as planning, retrospective and prospective memory, and visuospatial and verbal working memory, contribute to everyday multitasking. Subtasks were chosen to be diverse, and predictions were derived from a statistical model of everyday multitasking impairments associated with frontal-lobe lesions (Burgess, Veitch, de Lacy Costello, & Shallice, 2000b). Multiple regression indicated significant independent contributions from measures of retrospective memory, visuospatial working memory, and online planning, but not from independent measures of prospective memory or verbal working memory. Structural equation modelling showed that the best fit to the data arose from three underlying constructs, with Memory and Planning having a weak link, but with both having a strong directional pathway to an Intent construct that reflected implementation of intentions. Participants who followed their preprepared plan achieved higher scores than those who altered their plan during multitask performance. This was true regardless of whether the plan was efficient or poor. These results substantially develop and extend the Burgess et al. (2000b) model to healthy adults and yield new insight into the poorly understood area of everyday multitasking. The findings also point to the utility of using virtual environments for investigating this form of complex human cognition.

Time correlated anomaly detection based on inferences

Anomaly detection techniques are used to find the presence of anomalous activities in a network by comparing traffic data activities against a "normal" baseline. Although it has several advantages which include detection of "zero-day" attacks, the question surrounding absolute definition of systems deviations from its "normal" behaviour is important to reduce the number of false positives in the system. This study proposes a novel multi-agent network-based framework known as Statistical model for Correlation and Detection (SCoDe), an anomaly detection framework that looks for timecorrelated anomalies by leveraging statistical properties of a large network, monitoring the rate of events occurrence based on their intensity. SCoDe is an instantaneous learning-based anomaly detector, practically shifting away from the conventional technique of having a training phase prior to detection. It does acquire its training using the improved extension of Exponential Weighted Moving Average (EWMA) which is proposed in this study. SCoDe does not require any previous knowledge of the network traffic, or network administrators chosen reference window as normal but effectively builds upon the statistical properties from different attributes of the network traffic, to correlate undesirable deviations in order to identify abnormal patterns. The approach is generic as it can be easily modified to fit particular types of problems, with a predefined attribute, and it is highly robust because of the proposed statistical approach. The proposed framework was targeted to detect attacks that increase the number of activities on the network server, examples which include Distributed Denial of Service (DDoS) and, flood and flash-crowd events. This paper provides a mathematical foundation for SCoDe, describing the specific implementation and testing of the approach based on a network log file generated from the cyber range simulation experiment of the industrial partner of this project.

The influence of yarn parameters on the ultraviolet protection of yarns

 Improving ultraviolet (UV) protection of textiles is essential to protect wearers against UV radiation induced risks. In addition to fabric parameters, yarn parameters are important factors affecting UV protection of textiles. This work is to examine the influence of yarn parameters on UV protection in order to set up a statistical model for predicting the UV protection of yarns. Wool yarns with different variables were used to test the ultraviolet protection factor (UPF) values for data analysis and the model verification. The model provides the optimized parameters for the UV protective fabric design. This work is helpful as a pre-cursor to the development of a more advanced optical model, which will look at understanding the penetration of UV light through fibres, yarns and fabrics.

Measuring eco-roof mitigation on flash floods via GIS simulation

Purpose: The rapid and ongoing expansion of urbanised impervious areas could lead to more frequent flood inundation in urban flood-prone regions. Nowadays, urban flood inundation induced by rainstorm is an expensive natural disaster in many countries. In order to reduce the flooding risk, eco-roof systems (or green roof systems) could be considered as an effective mechanism of mitigating flooding disasters through their rainwater retention capability. However, there is still a lack of examining the stormwater management tool. The purpose of this paper is to evaluate the effects on flooding disaster from extensive green roofs. Design/methodology/approach: Based on geographical information system (GIS) simulation, this research presents a frame of assessing eco-roof impacts on urban flash floods. The approach addresses both urban rainfall-runoff and underground hydrologic models for traditional impervious and green roofs. Deakin University’s Geelong Waurn Ponds campus is chosen as a study case. GIS technologies are then utilised to visualise and analyse the effects on flood inundation from surface properties of building roofs. Findings: The results reveal that the eco-roof systems generate varying degrees of mitigation of urban flood inundation with different return period storms. Originality/value: Although the eco-roof technology is considered as an effective stormwater management tool, it is not commonly adopted and examined in urban floods. This study will bring benefits to urban planners for raising awareness of hazard impacts and to construction technicians for considering disaster mitigation via roof technologies. The approach proposed here could be used for the disaster mitigation in future urban planning.

A Bayesian approach to the determination of ignition delay

A novel in-cylinder pressure method for determining ignition delay has been proposed and demonstrated. This method proposes a new Bayesian statistical model to resolve the start of combustion, defined as being the point at which the band-pass in-cylinder pressure deviates from background noise and the combustion resonance begins. Further, it is demonstrated that this method is still accurate in situations where there is noise present. The start of combustion can be resolved for each cycle without the need for ad hoc methods such as cycle averaging. Therefore, this method allows for analysis of consecutive cycles and inter-cycle variability studies. Ignition delay obtained by this method and by the net rate of heat release have been shown to give good agreement. However, the use of combustion resonance to determine the start of combustion is preferable over the net rate of heat release method because it does not rely on knowledge of heat losses and will still function accurately in the presence of noise. Results for a six-cylinder turbo-charged common-rail diesel engine run with neat diesel fuel at full, three quarters and half load have been presented. Under these conditions the ignition delay was shown to increase as the load was decreased with a significant increase in ignition delay at half load, when compared with three quarter and full loads.

Evaluation of three daily rainfall generation models for SWAT

The Soil and Water Assessment Tool (SWAT) is a hydrologic model that was developed to predict the long-term impacts of land use change on the water balance of large catchments. Stochastic models are used to generate the daily rainfall sequences needed to conduct long-term, continuous simulations with SWAT. The objective of this study was to evaluate the performances of three daily rainfall generation models. The models evaluated were the modified Daily and Monthly Mixed (DMMm) model, skewed normal distribution (SKWD) model and modified exponential distribution (EXPD) model. The study area was the Woady Yaloak River catchment (306 km2) located in southwest Victoria, Australia. The models were assessed on their ability to preserve annual, monthly and daily statistical characteristics of the historical rainfall and runoff. The mean annual, monthly, and daily rainfall was preserved satisfactorily by the models. The DMMm model reproduced the standard deviation of annual and monthly rainfall better than the SKWD and EXPD models. Overall, the DMMm model performed marginally better than the SKWD model at reproducing the statistical characteristics of the historical rainfall record at the various time scales. The performance of the EXPD model was found to be inferior to the performances of the DMMm and SKWD models. The models reproduced the mean annual, monthly, and daily runoff relatively well, although the DMMm and SKWD models were found to preserve these statistics marginally better than the EXPD model. None of the models managed to reproduce the standard deviation of annual, monthly, and daily runoff adequately.

Productivity differential and the relationship between exports and GDP in Fiji : an empirical assessment using the two sector model

This study uses a two-sector model to determine the productivity differential between the export and non-export sectors of Fiji, and the contribution of exports and investment to gross domestic product over the period 1962-2000. Amongst our key results, we find that the productivity differential between the export and non-export sectors is small and statistically insignificant; investment to GDP ratio and weighted exports positively contribute to economic growth in Fiji; and in the abnormal years (years of coups in Fiji) marginal productivity in capital in the non-export sector is lower than in normal years.

An empirical model of international student satisfaction

Purpose – Based on the theoretical framework of expectancy-disconfirmation paradigm, the purpose of this paper is to examine the differences in student perceptions of the level of satisfaction related to educational and non-educational services among four groups of international postgraduate business students from China, India, Indonesia and Thailand undertaking study in Australia.

Design/methodology/approach – The data used in this study were derived from a mail survey conducted among international postgraduate business students from Asia studying at five universities in the state of Victoria, Australia. A total of 573 usable responses were received. Analysis using structural equation modelling, multivariate analysis of variance (MANOVA) and analysis of variance (ANOVA) was undertaken.

Findings – This study develops and tests a model of international postgraduate student satisfaction. Findings indicate that the importance of service quality factors related to both educational and non-educational services varies among nationality groups and, therefore, has a differential impact on student satisfaction.

Practical implications – The study provides insights into seven constructs related to educational and non-educational services that are perceived as important by postgraduate business students from Asia in satisfaction formation. Universities should develop a diversified strategic marketing plan that incorporates the differential needs of international postgraduate business students according to the educational and non-educational constructs developed in this paper.

Originality/value – This study makes a contribution by filling a void in academic research in the area of satisfaction in relation to postgraduate international business students from four nationality groups in Asia.

A model of price discovery and market design : theory and empirical evidence

An essential function of derivative markets is price discovery. A model is proposed to incorporate a comprehensive dynamic interaction between price size coordinates of orders and trades. An example of application of the model and its effect on price discovery is discussed.