Search

In this paper we describe our system for visualizing convection in the Earth's mantle using hybrid rendering - a combination of interactive raytracing, raycasting and conventional rasterization graphics. The dataset for the SciVis 2021 contest includes a velocity vector field and multiple scalar fields, such as temperature and thermal conductivity anomalies, which we display using isosurface raycasting, streamlines, and line integral convolution, among other techniques. We demonstrate how these methods can be used to explore the dataset, identify slabs, plumes and stagnations, and detect the influence of 660 km depth endothermic phase transition and mid-mantle spin transition on the mantle dynamics.

Mesh design is a major bottleneck in the creation of computer games and animation. Therefore simplifying the process of mesh editing is an important problem. Natural mesh processing is generally difficult to perform since natural meshes often have nontrivial surface shapes and high degrees of freedom. In this study I present novel approaches to natural mesh processing utilizing spectral methods based on the manifold harmonic basis. Geometric scaling transformations do not respect the biological processes which govern the size and shape of living creatures. I describe an approach to scaling which can be related to biological function. Known biological laws of allometry are used which are expressed as power laws to control the mesh deformation in the frequency domain. This approach is motivated by the relation between fractal biological systems and their underlying power-law spectra. I demonstrate my approach to biology-aware character scaling on triangle meshes representing quadrupedal mammals. Mesh editing is a time-consuming and error-prone process when changes must be manually applied to repeated structures in the mesh. I propose a fast and accurate method for performing region matching which is based on the manifold harmonic transform. Then this matching method is demonstrated in the context of nonlocal mesh editing - propagating mesh editing operations from a single source region to multiple target regions which may be arbitrarily far away. Finally, I show how mesh skeleton generation can be simplified using spectral approaches. These contributions will lead to more efficient methods of mesh editing and character design.

Renderings of 3D scenes can feature lines drawn automatically along sharp edges between colored areas on object textures, in order to imitate certain styles of hand-drawn line art. However, such “inter-color lines” have been studied very little. Two algorithms for rendering these lines were compared in this study—a faster one utilizing precomputed lines “baked” into the textures and a more complex one that dynamically generates the lines in image space on each frame—for the purpose of determining which of the two better imitated traditional, hand-drawn art styles and which was more visually appealing. In this study, test subjects compared results of the two algorithms side by side, both passively and interactively from a moving camera, and they noted the visual differences to reach final judgments as to which better adhered to artistic conventions and which was more appealing. Statistical analysis of the sample proportions that preferred each algorithm did not show that any significant difference existed between the two algorithms in terms of either of the above metrics. Thus the algorithm using precomputed lines appears to be more recommendable overall, as it is known to be computationally faster, whereas the dynamic algorithm was not shown to be preferred by viewers in terms of conventionality or aesthetics.

Fractal shapes reflect the behavior of complex natural systems, but can be generated by simple mathematical equations. Images of 3D fractals almost exclusively depict opaque surfaces, and use reflected light and shadows to simulate a physical realization of these virtual objects. But rich inner detail can be revealed by reinterpreting the fractal as a volume and considering material transparency, light absorption and refraction. This work explores the range of images made possible by employing volume rendering techniques inspired by medical image visualization.

Neuronal connectivity graphs can give valuable insight into the structure of the brain and patterns of functional activity. In this paper we describe a connectivity visualization technique using contrasting styles to emphasize the differences between pairs of graphs. Our proposed technique enables interactive exploration of connectivity graphs at the various levels of the hierarchical structure of the brain — atlas regions, brain lobes and cerebral hemispheres. We demonstrate our methods on data obtained from matched pairs of subjects in a Parkinson's disease (PD) study. Control group data is displayed in a smooth, organic style with continuous transitions between views at different levels of the brain hierarchy using edge bundling and a new node bundling technique. Data from the PD group is displayed with glitch and noise effects inspired by PD symptoms, highlighting the differences between each subject and their matched pair from the control group. We conclude by describing avenues for further evaluation of this preliminary work.

In this paper we describe our approach to the VAST Challenge Mini Challenge 3 which provided time-series multispectral image for visualization and analysis. We used Matlab for rapid prototyping, interface design and implementation of the final software. The resulting application, Multilab, allowed us to interactively explore the collection of multispectral images, generate and test hypotheses, and answer quantitative questions about the images. Index Terms: I.4.9 [Computing Methodologies]: Image Processing and Computer Vision—Applications; D.2.2 [Software]: Software Engineering–Design Tools and Techniques

In this work we describe an interactive technique which enables gestural curve and surface design in an immersive virtual environment. We use a pair of motion tracked controllers to allow the user to intuitively control a Hermite spline curve which can be swept through space to create surfaces, or to define the location and orientation of cloned meshes. A head mounted display and tracked controllers replace the traditional keyboard and mouse for view selection and object interaction. Natural and expressive body motions allow the user to specify the parameters which define the curves and surfaces. Results are demonstrated using the HTC Vive VR system.

Medical image atlases contain much information about human anatomy, but learning the shapes of anatomical regions and making sense of the overall structure defined in the atlas can be problematic. Atlases may contain hundreds of regions with complex shapes which can be tightly packed together. This makes visualisation difficult since the shapes can fit together in complex ways and visually obscure each other. In this work, we describe a technique which enables interactive exploration of medical image atlases that permits the hierarchical structure of the atlas and the content of an underlying medical image to be investigated simultaneously. Our method enables a user to create visualizations of the atlas similar to the exploded views used in technical illustrations to show the structure of mechanical assemblies. These views are constrained by the geometry of the atlas and the hierarchical structure to reduce the complexity of user interaction. We also enable the user to explode the atlas meshes themselves. The atlas meshes are registered with a medical image which is displayed on the cut surfaces of the meshes using raycasting. Results from the AAL human brain atlas are presented and discussed.

We present a method for faster and more flexible approximation of camera defocus effects given a focused image of a virtual scene and depth map. Our method leverages the advantages of low-rank linear filtering by reducing the problem of 2D convolution to multiple 1D convolutions, which significantly reduces the computational complexity of the filtering operation. In the case of rank 1 filters (e.g., the box filter and Gaussian filter), the kernel is described as `separable' since it can be implemented as a horizontal 1D convolution followed by a 1D vertical convolution. While many filter kernels which result in bokeh effects cannot be approximated closely by separable kernels, they can be effectively approximated by low-rank kernels. We demonstrate the speed and flexibility of low-rank filters by applying them to image blurring, tilt-shift postprocessing, and depth-of-field simulation, and also analyze the approximation error for several aperture shapes.

In this paper, we present a method for 3D mesh segmentation based on sparse non-negative matrix factorization (NMF). Image analysis techniques based on NMF have been shown to decompose images into semantically meaningful local features. Since the features and coefficients are represented in terms of non-negative values, the features contribute to the resulting images in an intuitive, additive fashion. Like spectral mesh segmentation, our method relies on the construction of an affinity matrix which depends on the geometric properties of the mesh. We show that segmentation based on the NMF is simpler to implement, and can result in more meaningful segmentation results than spectral mesh segmentation.

Geometric scaling transformations do not respect the biological processes which govern the size and shape of living creatures. In this paper, we describe an approach to scaling which can be related to biological function. We use known biological laws of allometry which are expressed as power laws to control the mesh deformation in the frequency domain. This approach is motivated by the relation between fractal biological systems and their underlying power-law spectra. We demonstrate our approach to biology-aware character scaling on triangle meshes representing quadrupedal mammals.