An Investigation Into Features For Multi-View Lipreading

For the first time in this paper we present results showing the effect of speaker head pose angle on automatic lip-reading performance over a wide range of closely spaced angles. We analyse the effect head pose has upon the features themselves and show that by selecting coefficients with minimum variance w.r.t. pose angle, recognition performance can be improved when train-test pose angles differ. Experiments are conducted using the initial phase of a unique multi view Audio-Visual database designed specifically for research and development of pose-invariant lip-reading systems. We firstly show that it is the higher order horizontal spatial frequency components that become most detrimental as the pose deviates. Secondly we assess the performance of different feature selection masks across a range of pose angles including a new mask based on Minimum Cross-Pose Variance coefficients. We report a relative improvement of 50% in Word Error Rate when using our selection mask over a common energy based selection during profile view lip-reading.

Feature Selection for Pose Invariant Lip Biometrics

For the first time in this paper the authors present results showing the effect of out of plane speaker head pose variation on a lip biometric based speaker verification system. Using appearance DCT based features, they adopt a Mutual Information analysis technique to highlight the class discriminant DCT components most robust to changes in out of plane pose. Experiments are conducted using the initial phase of a new multi view Audio-Visual database designed for research and development of pose-invariant speech and speaker recognition. They show that verification performance can be improved by substituting higher order horizontal DCT components for vertical, particularly in the case of a train/test pose angle mismatch.

POINTS OF VIEW: Multi-Roling Performance in Ireland in the 1990s, Part 1 (Layering)

This series of posts will consider the acting techniques required for this style of performance with reference to four productions from the 1990s in which two actors took on multiple roles: Frank Pig Says Hello (1992), Co-Motion’s stage adaptation of Pat McCabe’s novel; The Butcher Boy, Corca Dorca’s production of Disco Pigs (1996) by Enda Walsh; DubbleJoint’s production of Stones in his Pockets by Marie Jones (1996) and its subsequent award-winning revival which I produced at Belfast’s Lyric Theatre in 1999; and Kabosh’s production of Mojo Mickybo by Owen McCafferty (1998). The discussion will focus primarily on my own empirical exploration of the demands multi-roling places on an actor through the direction of recent revivals of Mojo Mickybo for Belfast’s Chatterbox Theatre Company (2013) and Bedlam Productions (2015).

POINTS OF VIEW: Multi-Roling Performance in Ireland in the 1990s, Part 2 (Characterisation)

The 1990s in Ireland saw a series of highly successful theatre productions in which actors played a multiplicity of roles. This has often been attributed to the economic exigencies of the times, but it also depended on the availability of flexible actors with the physical and psychological capacity to embody a wide range of identifiable characters within the one production.

This second of two posts considers the acting techniques required for this style of performance in relation to the differentiation of one character from another. The discussion will focus primarily on my own empirical exploration of the demands multi-roling places on an actor through the direction of recent revivals of Mojo Mickybo for Belfast’s Chatterbox Theatre Company (2013) and Bedlam Productions (2015).

Atomic excitation during recollision-free ultrafast multi-electron tunnel ionization

Modern intense ultrafast pulsed lasers generate an electric field of sufficient strength to permit tunnel ionization of the valence electrons in atoms(1). This process is usually treated as a rapid succession of isolated events, in which the states of the remaining electrons are neglected(2). Such electronic interactions are predicted to be weak, the exception being recollision excitation and ionization caused by linearly polarized radiation(3). In contrast, it has recently been suggested that intense field ionization may be accompanied by a two-stage 'shake-up' reaction(4). Here we report a unique combination of experimental techniques(5-8) that allows us to accurately measure the tunnel ionization probability for argon exposed to 50-fs laser pulses. Most significantly for the current study, this measurement is independent of the optical focal geometry(7,8), equivalent to a homogenous electric field. Furthermore, circularly polarized radiation negates recollision. The present measurements indicate that tunnel ionization results in simultaneous excitation of one or more remaining electrons through shake-up(9). From an atomic physics standpoint, it may be possible to induce ionization from specific states, and will influence the development of coherent attosecond extreme-ultraviolet-radiation sources(10). Such pulses have vital scientific and economic potential in areas such as high-resolution imaging of in vivo cells and nanoscale extreme-ultraviolet lithography.

On the recollision-free excitation of krypton during ultrafast multi-electron tunnel ionization

The probability of multiple ionization of krypton by 50 femtosecond circularly polarized laser pulses, independent of the optical focal geometry, has been obtained for the first time. The excellent agreement over the intensity range 100 TW cm-2 to 100 PW cm-2 with the recent predictions of Kornev et al (2003 Phys. Rev. A 68 043414) provides the first experimental confirmation that non-recollisional electronic excitation can occur in strong-field ionization. This is particularly true for higher stages of ionization, when the laser intensity exceeds 10 PW cm-2 as the energetic departure of the ionized electron(s) diabatically distorts the wavefunctions of the bound electrons. By scaling the probability of ionization by the focal volume, we discuss why this mechanism was not apparent in previous studies.

A constraints-based resource discovery model for multi-provider cloud environments

Increasingly infrastructure providers are supplying the cloud marketplace with storage and on-demand compute resources to host cloud applications. From an application user's point of view, it is desirable to identify the most appropriate set of available resources on which to execute an application. Resource choice can be complex and may involve comparing available hardware specifications, operating systems, value-added services, such as network configuration or data replication, and operating costs, such as hosting cost and data throughput. Providers' cost models often change and new commodity cost models, such as spot pricing, have been introduced to offer significant savings. In this paper, a software abstraction layer is used to discover infrastructure resources for a particular application, across multiple providers, by using a two-phase constraints-based approach. In the first phase, a set of possible infrastructure resources are identified for a given application. In the second phase, a heuristic is used to select the most appropriate resources from the initial set. For some applications a cost-based heuristic is most appropriate; for others a performance-based heuristic may be used. A financial services application and a high performance computing application are used to illustrate the execution of the proposed resource discovery mechanism. The experimental result shows the proposed model could dynamically select an appropriate set of resouces that match the application's requirements.

Organization specific predictors of job satisfaction: findings from a Canadian multi-site quality of work life cross-sectional survey

Background: Organizational features can affect how staff view their quality of work life. Determining staff perceptions about quality of work life is an important consideration for employers interested in improving employee job satisfaction. The purpose of this study was to identify organization specific predictors of job satisfaction within a health care system that consisted of six independent health care organizations.

Methods: 5,486 full, part and causal time (non-physician) staff on active payroll within six organizations (2 community hospitals, 1 community hospital/long-term care facility, 1 long-term care facility, 1 tertiary care/community health centre, and 1 visiting nursing agency) located in five communities in Central West Ontario, Canada were asked to complete a 65-item quality of work life survey. The self-administered questionnaires collected staff perceptions of: co-worker and supervisor support; teamwork and communication; job demands and decision authority; organization characteristics; patient/resident care; compensation and benefits; staff training and development; and impressions of the organization. Socio-demographic data were also collected.

Results: Depending on the organization, between 15 and 30 (of the 40 potential predictor) variables were found to be statistically associated with job satisfaction (univariate analyses). Logistic regression analyses identified the best predictors of job satisfaction and these are presented for each of the six organizations and for all organizations combined.

Conclusions: The findings indicate that job satisfaction is a multidimensional construct and although there appear to be some commonalities across organizations, some predictors of job satisfaction appear to be organization and context specific.

Conjugate heat transfer and film cooling of a multi-stage cooling scheme

Cooling techniques play a key role in improving efficiency and power output of modern gas turbines. The conjugate technique of film and impingement cooling schemes is considered in this study. The Multi-Stage Cooling Scheme (MSCS) involves coolant passing from inside to outside turbine blade through two stages. The first stage; the coolant passes through first hole to internal gap where the impinging jet cools the external layer of the blade. Finally, the coolant passes through the internal gap to the second hole which has specific designed geometry for external film cooling. The effect of design parameters, such as, offset distance between two-stage holes, gap height, and inclination angle of the first hole, on upstream conjugate heat transfer rate and downstream film cooling effectiveness performance are investigated computationally. An Inconel 617 alloy with variable properties is selected for the solid material. The conjugate heat transfer and film cooling characteristics of MSCS are analyzed across blowing ratios of Br = 1 and 2 for density ratio, 2. This study presents upstream wall temperature distributions due to conjugate heat transfer for different gap design parameters. The maximum film cooling effectiveness with upstream conjugate heat transfer is less than adiabatic film cooling effectiveness by 24–34%. However, the full coverage of cooling effectiveness in spanwise direction can be obtained using internal cooling with conjugate heat transfer, whereas adiabatic film cooling effectiveness has narrow distribution.

The effect of nozzle geometry on the flow characteristics of small water jets

A wide variety of processes make use of plain orifice nozzles. Fuel injectors for internal combustion engines incorporate these nozzles to generate finely atomized sprays. Processes such as jet cutting, jet cleaning, and hydroentanglement, on the other hand, use similar nozzles, but require coherent jets. The spray or jet characteristics depend on the stability of the flow emerging from the orifice. This problem has been extensively researched for nozzles with diameters above 300 μm. Much less is known about the characteristics of jets produced by nozzles with smaller diameters, where viscous effects and small geometric variations due to manufacturing tolerances are likely to play an increasing role. Results are presented of a wide-ranging investigation of geometry effects on the flow parameters and jet characteristics of nozzles with diameters between 120 and 170 μm. Nozzles with circular cross-section and conical, cone-capillary and capillary axial designs were investigated. For conical and cone-capillary nozzles, the effect of cone angle and effects due to interactions between adjacent nozzles in the multi-hole cone-capillary nozzles were studied. For capillary nozzles, the effects of diameter variations and inlet edge roundness for capillary nozzles were considered. Furthermore, the effect of varying the aspect ratio (ratio of major and minor axes) of elliptical nozzles was studied. Flowrate and jet impact force measurements were carried out to determine the discharge coefficient C, velocity coefficient C, and contraction coefficient C of the nozzles for supply pressures between 3 and 12 MPa. Visualizations of the jet flow were carried out in the vicinity of the nozzle exit in order to identify near-nozzle flow regimes and to study jet coherence. The relationship between nozzle geometry, discharge characteristics, and jet coherence is examined. © IMechE 2006.

Effect of nozzle geometry on the flow characteristics of hydroentangling jets

Cone-capillary nozzles with varying cone angles from 10° to 120° and a capillary diameter of 120μ are experimentally investigated for their application in the hydroentanglement process. Cone-up and cone-down configurations in a range of water pressures of 30-120 bar are tested. The effects of the cone angle on flow parameters such as discharge and velocity coefficients and intact length are studied. Flow visualization techniques are used to recognize the flow regimes and characteristics and to inspect and compare the intact length and appearance of the jets. Cone-down nozzles with more consistent flow properties, lower discharges, and higher velocity coefficients are more suitable for the hydroentanglement process. Single-cone nozzles without capillaries and with varying cone angles are also tested. The flow properties of the jets from the single-cone nozzles are compared with the cone-capillary nozzles of the same cone angle to study the effect of the capillary section. The effect of the interaction of adjacent nozzles on the flow from multi-hole nozzles is studied, and the characteristics of the jets from the multi-hole nozzles are compared with the single-hole nozzles.

Constrained Multi-Objective Aerofoil Design Using a Multi-Level Optimisation Strategy

A novel approach for the multi-objective design optimisation of aerofoil profiles is presented. The proposed method aims to exploit the relative strengths of global and local optimisation algorithms, whilst using surrogate models to limit the number of computationally expensive CFD simulations required. The local search stage utilises a re-parameterisation scheme that increases the flexibility of the geometry description by iteratively increasing the number of design variables, enabling superior designs to be generated with minimal user intervention. Capability of the algorithm is demonstrated via the conceptual design of aerofoil sections for use on a lightweight laminar flow business jet. The design case is formulated to account for take-off performance while reducing sensitivity to leading edge contamination. The algorithm successfully manipulates boundary layer transition location to provide a potential set of aerofoils that represent the trade-offs between drag at cruise and climb conditions in the presence of a challenging constraint set. Variations in the underlying flow physics between Pareto-optimal aerofoils are examined to aid understanding of the mechanisms that drive the trade-offs in objective functions.

A numerical study of trailing edge geometry for low inertia automotive mixed flow turbines

Abstract. Mixed flow turbines can offer improvements over typical radial turbines used in automotive turbochargers, with respect to transient performance and low velocity ratio efficiency. Turbine rotor mass dominates the rotating inertia of the turbocharger’s rotating assembly, and any reductions of mass in the outer radii of the wheel, including the rotor back-disk, can significantly reduce this inertia and improve the acceleration of the assembly. Off-design, low velocity ratio conditions are typified by highly tangential flow at the rotor inlet and a non-zero inlet blade angle is desirable for such operating conditions. This is achievable in a Mixed Flow Turbine without increasing bending stresses within the rotor blade, which is beneficial in high speed and high inlet temperature turbine designs.
This study considers the meridional geometry of Mixed Flow Turbines using a multi-disciplinary study to assess both the structural and aerodynamic performance of each rotor, incorporating both CFD and FEA. Variations of rotor trailing edge were investigated at different operating conditions representing both on- and off-design operation within the constraints of existing hardware geometries. In all cases, the performance is benchmarked against an existing state-of-the-art radial turbocharger turbine with consideration of rotor inertia and its benefit for engine transient performance. The results indicate the influence of these parameters and this report details their benefits with respect to turbocharging a downsized, automotive engine.