An Invariant Three-polar Representation for R3 Surfaces: Robustness and Accuracy for 3D Faces Description

In this paper, we intend to introduce a new curved surface representation that we qualify by three-polar. It is constructed by the superposition of the three geodesic potentials generated from three reference points of the surface. By considering a pre-selected levels set of this superposition, invariant points are obtained. The accuracy of the three-polar representation for 3D human faces description is performed in the mean of the Hausdorff distance. A comparison between this representation and the one based on the level curves around the nose tip is established in the sense of the robustness under errors on the nose tip positions.

Innovative tooth contact analysis with non-uniform rational b-spline surfaces

More and more simulation tools are being used in the development of gears in order to save development time and costs while improving the gears. BECAL is a comprehensive software tool for the tooth contact analysis (TCA) of bevel, hypoid, beveloid and spur gears. The gear geometry is provided by a manufacturing simulation or a geometry import. To determine the exact contact conditions in the TCA, the discrete flank points are converted into a continuous and differentiable surface representation. At present, it is an approximation by means of Bézier tensor product surfaces. With this surface representation, significant deviations to the target points can occur depending on the tooth geometry. In particular tip, root and end relief, strongly curved tooth root geometries or discontinuous topological measurement data due to e.g. micro-pitting can only be considered insufficiently.Hence, a new method for surface approximation with non-uniform rational b‑spline surfaces (NURBS) is presented. Its application can significantly improve the surface representation compared to the target geometry, leading to more realistic results regarding contact stress, tooth root stress and transmission error. To illustrate the advantages, NURBS-based surfaces are compared with the Bézier tensor product surfaces. Finally, the potential of the new approach regarding the prediction of lifetime and acoustics is demonstrated by application to different gear geometries.

Surface representation impacts on turbulent heat fluxes in the Weather Research and Forecasting (WRF) model (v.4.1.3)

The water and energy transfers at the interface between the Earth's surface and the atmosphere should be correctly simulated in numerical weather and climate models. This implies the need for a realistic and accurate representation of land cover (LC), including appropriate parameters for each vegetation type. In some cases, the lack of information and crude representation of the surface lead to errors in the simulation of soil and atmospheric variables. This work investigates the ability of the Weather Research and Forecasting (WRF) model to simulate surface heat fluxes in a heterogeneous area of southern France using several possibilities for the surface representation. In the control experiments, we used the default LC database in WRF, which differed significantly from the actual LC. In addition, sub-grid variability was not taken into account since the model uses, by default, only the surface information from the dominant LC category in each pixel (dominant approach). To improve this surface simplification, we designed three new interconnected numerical experiments with three widely used land surface models (LSMs) in WRF. The first one consisted of using a more realistic and higher-resolution LC dataset over the area of analysis. The second experiment aimed at investigating the effect of using a mosaic approach; 30 m sub-grid surface information was used to calculate the final grid fluxes based on weighted averages from values obtained for each LC category. Finally, in the third experiment, we increased the model stomatal conductance for conifer forests due to the large flux errors associated with this vegetation type in some LSMs. The simulations were evaluated with gridded area-averaged fluxes calculated from five tower measurements obtained during the Boundary-Layer Late Afternoon and Sunset Turbulence (BLLAST) field campaign. The results from the experiments differed depending on the LSM and displayed a high dependency of the simulated fluxes on the specific LC definition within the grid cell, an effect that was enhanced with the dominant approach. The simulation of the fluxes improved using the more realistic LC dataset except for the LSMs that included extreme surface parameters for coniferous forest. The mosaic approach produced fluxes more similar to reality and served to particularly improve the latent heat flux simulation of each grid cell. Therefore, our findings stress the need to include an accurate surface representation in the model, including soil and vegetation sub-grid information with updated surface parameters for some vegetation types, as well as seasonal and man-made changes. This will improve the modelled heat fluxes and ultimately yield more realistic atmospheric processes in the model.

Short Association Fibre Tractography

Through advancing the existing and introducing novel methodological developments in streamlines tractography, this work proposes an approach that is meant to specifically interrogate an important yet relatively understudied population of the human white matter - the short association fibres. By marrying tractography with surface representation of the cortex, the framework: (1) ensures a greater cortical surface coverage through spreading streamline seeds more uniformly; (2) relies on precise filtering mechanics which are particularly important when dealing with small, morphologically complex structures; (3) allows to make use of surface-based registration for dataset comparisons which can be superior in the vicinity of the cortex. The indexation of surface vertices at each streamline end enables direct interfacing between streamlines and the cortical surface without dependence on the voxel grid. Short association fibre tractograms generated using recent test-retest data from our institution are carefully characterised and measures of consistency using streamline-, voxel-, surface- and network-wise comparisons calculated.

WHITE-BOX CARTOONIZATION USING AN EXTENDED GAN FRAMEWORK

In the present study, we propose to implement a new framework for estimating generative models via an adversarial process to extend an existing GAN framework and develop a white-box controllable image cartoonization, which can generate high-quality cartooned images/videos from real-world photos and videos. The learning purposes of our system are based on three distinct representations: surface representation, structure representation, and texture representation. The surface representation refers to the smooth surface of the images. The structure representation relates to the sparse colour blocks and compresses generic content. The texture representation shows the texture, curves, and features in cartoon images. Generative Adversarial Network (GAN) framework decomposes the images into different representations and learns from them to generate cartoon images. This decomposition makes the framework more controllable and flexible which allows users to make changes based on the required output. This approach overcomes any previous system in terms of maintaining clarity, colours, textures, shapes of images yet showing the characteristics of cartoon images.

Surface representation impacts on turbulent heat fluxes in WRF (v.4.1.3)

The water and energy transfers at the interface between the Earth's surface and the atmosphere should be correctly simulated in numerical weather and climate models. This implies the need for a realistic and accurate representation of land cover (LC), including appropriate parameters for each vegetation type. In some cases, the lack of information and crude representation of the surface leads to errors in the simulation of soil and atmospheric variables. This work investigates the ability of the Weather Research and Forecasting (WRF) model to simulate surface heat fluxes in a heterogeneous area of southern France, using several possibilities for the surface representation. In the control experiments, we used the default LC database in WRF, which differed significantly from the actual LC. In addition, sub-grid variability was not taken into account since the model uses, by default, only the surface information from the dominant LC category in each pixel (dominant approach). To improve this surface simplification, we designed three new interconnected numerical experiments with four widely-used land-surface models (LSMs) in WRF. The first one consisted of using a more realistic and higher-resolution LC dataset over the area of analysis. The second experiment aimed at investigating the effect of using a mosaic approach, where 30-m sub-grid surface information was used to calculate the final grid fluxes, based on weighted averages from values obtained for each LC category. Finally, in the third experiment, we increased the model stomatal conductance for conifer forests, due to the large fluxes errors associated with this vegetation type in some LSMs. The simulations were evaluated with gridded area-averaged fluxes calculated from five tower measurements obtained during the Boundary Layer Late Afternoon and Sunset Turbulence (BLLAST) field campaign. The results from the experiments differed depending on the LSM and displayed a high dependency of the simulated fluxes on the specific LC definition within the grid cell, an effect that was enhanced with the dominant approach. The simulation of the fluxes improved using the more realistic LC dataset except for the LSMs that included extreme surface parameters for the coniferous forest. The mosaic approach produced fluxes more similar to reality and served to improve, especially, the latent heat flux simulation of each grid cell. Therefore, our findings stress the need to include an accurate surface representation in the model, including soil and vegetation sub-grid information with updated surface parameters for some vegetation types, as well as seasonal and man-made changes. This will improve the modelled heat fluxes and ultimately yield more realistic atmospheric processes in the model.