Automatic and Efficient Human Pose Estimation for Sign Language Videos

We present a fully automatic arm and hand tracker that detects joint positions over continuous sign language video sequences of more than an hour in length. To achieve this, we make contributions in four areas: (i) we show that the overlaid signer can be separated from the background TV broadcast using co-segmentation over all frames with a layered model; (ii) we show that joint positions (shoulders, elbows, wrists) can be predicted per-frame using a random forest regressor given only this segmentation and a colour model; (iii) we show that the random forest can be trained from an existing semi-automatic, but computationally expensive, tracker; and, (iv) introduce an evaluator to assess whether the predicted joint positions are correct for each frame. The method is applied to 20 signing footage videos with changing background, challenging imaging conditions, and for different signers. Our framework outperforms the state-of-the-art long term tracker by Buehler et al. (International Journal of Computer Vision 95:180–197, 2011), does not require the manual annotation of that work, and, after automatic initialisation, performs tracking in real-time. We also achieve superior joint localisation results to those obtained using the pose estimation method of Yang and Ramanan  (Proceedings of the IEEE conference on computer vision and pattern recognition, 2011).     

Effect of Grain Size on Mechanical Properties of Submicrometer 3Y-TZP: Fracture Strength and Hydrothermal Degradation

This paper considers fracture strength, fracture origins, and hydrothermal degradation of 3Y-TZP with grain sizes in the range of 110–480 nm. Biaxial fracture strength testing was used to show that the fracture strength increases with grain size and is governed by the concurrent change in fracture toughness. Hydrothermal degradation was studied by means of fractography, Raman microscopy and its effect on fracture strength. Up to 200 nm grain size, hydrothermal degradation of strength is limited. Larger grain sizes exhibit either premature failure or an increase in strength. A surface transformation zone was found to be responsible for both phenomena.     

Highly Transparent Lu-α-SiAlON

Novel Lu-α-SiAlON ceramics were produced by hot pressing mixtures of Si₃N₄, Lu₂O₃, AlN, and Al₂O₃ at 1950°C for 2 h in a nitrogen atmosphere. The resultant SiAlON was fully dense and possessed a uniform, equiaxed microstructure with a grain size of ∼1 μm, which resulted in a high hardness of >19 GPa. In addition to high hardness, the sample showed very high optical transparency in the visible light region, with >70% transmission at higher wavelengths. This high transparency was attributed to the uniform, dense microstructure and lack of residual grain-boundary phase.     

Interactional Order and Constructed Ways of Seeing with Touchless Imaging Systems in Surgery

While surgical practices are increasingly reliant on a range of digital imaging technologies, the ability for clinicians to interact and manipulate these digital representations in the operating theatre using traditional touch based interaction devices is constrained by the need to maintain sterility. To overcome these concerns with sterility, a number of researchers are have been developing ways of enabling interaction in the operating theatre using touchless interaction techniques such as gesture and voice to allow clinicians control of the systems. While there have been important technical strides in the area, there has been little in the way of understanding the use of these touchless systems in practice. With this in mind we present a touchless system developed for use during vascular surgery. We deployed the system in the endovascular suite of a large hospital for use in the context of real procedures. We present findings from a study of the system in use focusing on how, with touchless interaction, the visual resources were embedded and made meaningful in the collaborative practices of surgery. In particular we discuss the importance of direct and dynamic control of the images by the clinicians in the context of talk and in the context of other artefact use as well as the work performed by members of the clinical team to make themselves sensable by the system. We discuss the broader implications of these findings for how we think about the design, evaluation and use of these systems.     

The technology of halogen-free flame retardant phosphorus additives for polymeric systems

Halogen-free flame retardant additives based on phosphorus function by developing a protective char. A variety of additives are available, ranging from the element itself, in the form of amorphous red phosphorus, to specialty organophosphorus compounds, and examples of their use in a range of thermoplastics are given. Intumescent formulations based on phosphates are especially designed for polyolefins. The behavior of a typical intumescent system is described with respect to flame retardant performance, thermal stability, water sensitivity, and filler compatibility.     

A polyamide-silica composite prepared by the sol-gel process

Summary A new kind of organic-inorganic hybrid composite was prepared by means of the sol-gel process. The polymer employed was a mixed-isomer aromatic polyamide having good solubility and thermal stability. The silica constituting the inorganic phase was produced by the hydrolysis and condensation of tetramethoxysilane. The bonding between the phases involved aminophenyl-trimethoxysilane, in which the amino group can react with the phthaloyl chloride end-capped polymer, and the methoxysilane groups undergo hydrolysis. The composition of these composites was varied by changing the linear polymer chain length and relative amount of tetramethoxysilane. The gelation time was found to range from a few minutes to several days. Thermogravimetric analyses showed that decomposition starts at approximately 450 °C. Thin films cast from materials having a relatively high silica content were opaque and rigid, but those with low silica content were flexible and transparent.     

Characterisation of the adsorption sites of hydrogen on Pt/C fuel cell catalysts

A key component of a hydrogen fuel cell is a catalyst to dissociate dihydrogen to hydrogen atoms. In the present study, the adsorption of hydrogen on Pt/C fuel cell catalysts has been investigated by inelastic neutron scattering spectroscopy.

Monitoring a clean Pt(50%)/C catalyst with low energy neutron spectroscopy, after exposure to dihydrogen at 20 K, as it was heated to room temperature, showed three distinct temperature regimes: (i) a decrease in intensity from 10 to 60 K, (ii) a rise to a maximum between 60 and 120 K and then (iii) a slow fall-off towards room temperature. We assign the three regions as: (i) desorption of physisorbed dihydrogen, (ii) dissociation of dihydrogen to give an adsorbed layer and (iii) damping of the response by an increasing Debye–Waller factor.

The vibrational INS spectra of a series of Pt/C catalysts prepared under varying conditions were similar indicating that the same types of site are common to all the catalysts, although the relative proportions of each site are sample dependent. Features at 520, 950 and part of the intensity at 1300 cm⁻¹ are assigned to hydrogen on (1 1 1) faces, in good agreement with single crystal data. The mode at 640 cm⁻¹ is assigned as the doubly degenerate asymmetric stretch of Pt(1 0 0) faces with the symmetric stretch near 550 cm⁻¹.

We assign the bending mode of the on-top site to the feature at 470 cm⁻¹. The Pt–H stretch mode was observed at 2079 cm⁻¹. This is a significant result: this is the first time that hydrogen on the on-top sites has been observed on nanosized platinum particles supported on high surface area carbon black. The width of the INS peak is surprisingly large and may give additional information on the type and relative proportions of the crystallographic faces present on the catalyst particles.     

Dynamic systems modelling using genetic programming

In this contribution genetic programming (GP) is used to evolve dynamic process models. An innovative feature of the GP algorithm is its ability to automatically discover the appropriate time history of model terms required to build an accurate model. Two case studies are used to compare the performance of the GP algorithm with that of filter-based neural networks (FBNNs). Although the models generated using GP have comparable prediction performance to the FBNN models, a disadvantage is that they required greater computational effort to develop. However, we show that a major benefit of the GP approach is that additional model performance criteria can be included during the model development process. The parallel nature of GP means that it can evolve a set of candidate solutions with varying levels of performance in each objective. Although any combination of model performance criteria could be used as objectives within a multi-objective GP (MOGP) framework, the correlation tests outlined by Billings and Voon (Int. J. Control 44 (1986) 235) were used in this work.     

Exploiting border structure in solving large sparse linear systems in bordered triangular form

An elimination strategy for solving sparse linear systems in bordered triangular form is presented. This strategy is designed to exploit certain common structural properties of the borders of such a system. It may also be interpreted as a recursive application of block elimination. The technique is numerically efficient and provides significant reductions in fill-in compared to routine Gaussian elimination. It is applicable in the iterative solution of the large systems of nonlinear equations that arise in the equation-oriented approach to process simulation and design calculations, as well as in other problems of interest to the chemical engineer.