Combination rules in the Dempster-Shafer theory aim to summarize multiple corpuses of evidence that come from different sources. However, these summarizations are computationally demanding as they usually require work...
详细信息
Combination rules in the Dempster-Shafer theory aim to summarize multiple corpuses of evidence that come from different sources. However, these summarizations are computationally demanding as they usually require working with large amounts of information, which prevents their use in real life problems. In this work, different algorithms are proposed and compared in order to determine the fastest techniques to combine information under the Dempster-Shafer theory framework. These algorithms are created for Dempster's original combination rule and also for other modifications of this rule. Also, functions for combining sources using averaging combination rules are provided. The algorithms proposed in this work are designed to be executed in a Graphical Processing Unit (gpu) and have been implemented using Python and CUDA. The use of a gpu, which can execute multiple tasks in parallel, makes the algorithms faster than classic algorithms developed to be executed in a CPU. Results show the feasibility of the implementations proposed in this work that, using Python and CUDA, are able to combine corpuses of evidence for frames of discernment up to 28 elements in seconds.
When rendering light colored hair, multiple fiber scattering is essential for the right perception of the overall hair color. In this context, we present a novel technique to efficiently approximate multiple fiber sca...
详细信息
When rendering light colored hair, multiple fiber scattering is essential for the right perception of the overall hair color. In this context, we present a novel technique to efficiently approximate multiple fiber scattering for a full head of human hair or a similar fiber based geometry. In contrast to previous ad-hoc approaches, our method relies on the physically accurate concept of the Bidirectional Scattering Distribution Functions and gives physically plausible results with no need for parameter tweaking. We show that complex scattering effects can be approximated very well by using aggressive simplifications based on this theoretical model. When compared to unbiased Monte-Carlo path tracing, our approximations preserve photo-realism in most settings but with rendering times at least two-orders of magnitude lower. Time and space complexity are much lower compared to photon mapping-based techniques and we can even achieve realistic results in real-time on a standard PC with consumer graphics hardware.
The standard bilinear interpolation on normal maps results in visual artifacts along sharp features, which are common for surfaces with creases, wrinkles, and dents. In many cases, spatially varying features, like the...
详细信息
The standard bilinear interpolation on normal maps results in visual artifacts along sharp features, which are common for surfaces with creases, wrinkles, and dents. In many cases, spatially varying features, like the normals near discontinuity curves, are best represented as functions of the distance to the curve and the position along the curve. For high-quality interactive rendering at arbitrary magnifications, one needs to interpolate the distance field preserving discontinuity curves exactly. We present a real-time, gpu-based method for distance function and distance gradient interpolation which preserves discontinuity feature curves. The feature curves are represented by a set of quadratic Bezier curves, with minimal restrictions on their intersections. We demonstrate how this technique can be used for real-time rendering of complex feature patterns and blending normal maps with procedurally defined profiles near normal discontinuities.
We present a Minkowski sum algorithm for polyhedra based on convolution. We develop robust CPU and gpu implementations, using our ACP strategy to eliminate degeneracy and to enforce a user-specified backward error bou...
详细信息
We present a Minkowski sum algorithm for polyhedra based on convolution. We develop robust CPU and gpu implementations, using our ACP strategy to eliminate degeneracy and to enforce a user-specified backward error bound. We test the programs on 45 inputs with an error bound of 10(-8). The CPU program outperforms prior work, including non-robust programs. The gpu program using 2688 CUDA cores exhibits a median speedup factor of 36, which increases to 68 on the 6 hardest tests. For example, it computes a Minkowski sum with a million features in 20 seconds. (C) 2015 Elsevier Ltd. All rights reserved.
We present a new method for the real-time simulation of fluid surface waves and their interactions with floating objects. The method is based on the new concept of wave particles, which offers a simple, fast, and unco...
详细信息
We present a new method for the real-time simulation of fluid surface waves and their interactions with floating objects. The method is based on the new concept of wave particles, which offers a simple, fast, and unconditionally stable approach to wave simulation. We show how graphics hardware can be used to convert wave particles to a height field surface, which is warped horizontally to account for local wave-induced flow. The method is appropriate for most fluid simulation situations that do not involve significant global flow. It is demonstrated to work well in constrained areas, including wave reflections off of boundaries, and in unconstrained areas, such as an ocean surface. Interactions with floating objects are easily integrated by including wave forces on the objects and wave generation due to object motion. Theoretical foundations and implementation details are provided, and experiments demonstrate that we achieve plausible realism. Timing studies show that the method is scalable to allow simulation of wave interaction with several hundreds of objects at real-time rates.
Ruled surfaces play an important role in many manufacturing and construction applications. In this work, we explore a multi-dimensional dynamic programming based ruled surface fitting scheme to a given freeform ration...
详细信息
Ruled surfaces play an important role in many manufacturing and construction applications. In this work, we explore a multi-dimensional dynamic programming based ruled surface fitting scheme to a given freeform rational surface, S. Considering two initial opposite boundaries of S, sampled into a discrete piecewise linear polyline representation, the ruled surface fitting problem is reduced to a pairing-search between the polylines and elevations above the polylines, in the normal directions of S. A four-dimensional dynamic programming solution is sought for the four dimensions prescribed by the two polylines and the two elevation levels along the surface normals. This multi-dimensional dynamic programming is evaluated using highly parallel algorithms running on gpus that ensures the best fit to the sampled data. In order to evaluate the fitting error with respect to S, we derive a scheme to compute a bound from above on the maximal error between a bilinear surface patch (formed by two consecutive point-pairs) and its corresponding surface region on S. Surface-surface composition is employed to extract the corresponding surface region on S to compare against. Finally, the above ruled surface fitting approach is also extended into a discrete algorithm to find the non-isoparametric subdivision curve on S when a discrete recursive piecewise-ruled surface fitting is considered. A five- or seven-dimensional dynamic programming solution is employed towards this end and once again, surface-surface composition is employed to extract the two subdivided patches as tensor products. (C) 2014 Elsevier Ltd. All rights reserved.
Caustics are crucial in water rendering, yet they are often neglected in real-time applications due to the demanding computational requirements of the general-purpose caustics computation methods. In this paper we pre...
详细信息
Caustics are crucial in water rendering, yet they are often neglected in real-time applications due to the demanding computational requirements of the general-purpose caustics computation methods. In this paper we present a two-pass algorithm for caustics computation that is extremely fast and produces high-quality results. Our algorithm is targeted for commonly used height field representations of water and a planar caustic-receiving surface. The underlying theory of our approach is presented along with implementation details and pseudo codes.
Medical simulation frameworks facilitate both the preoperative and postoperative analysis of the patient's pathophysical condition. Of particular importance is the simulation of radiation dose delivery for real-ti...
详细信息
Medical simulation frameworks facilitate both the preoperative and postoperative analysis of the patient's pathophysical condition. Of particular importance is the simulation of radiation dose delivery for real-time radiotherapy monitoring and retrospective analyses of the patient's treatment. In this paper, a software framework tailored for the development of simulation-based real-time radiation dose monitoring medical applications is discussed. A multi-gpu-based computational framework coupled with inter-process communication methods is introduced for simulating the radiation dose delivery on a deformable 3D volumetric lung model and its real-time visualization. The model deformation and the corresponding dose calculation are allocated among the gpus in a task-specific manner and is performed in a pipelined manner. Radiation dose calculations are computed on two different gpu hardware architectures. The integration of this computational framework with a front-end software layer and back-end patient database repository is also discussed. Real-time simulation of the dose delivered is achieved at once every 120 ms using the proposed framework. With a linear increase in the number of gpu cores, the computational time of the simulation was linearly decreased. The inter-process communication time also improved with an increase in the hardware memory. Variations in the delivered dose and computational speedup for variations in the data dimensions are investigated using D70 and D90 as well as gEUD as metrics for a set of 14 patients. Computational speed-up increased with an increase in the beam dimensions when compared with a CPU-based commercial software while the error in the dose calculation was < 1%. Our analyses show that the framework applied to deformable lung model-based radiotherapy is an effective tool for performing both real-time and retrospective analyses.
Colocation patterns refer to subsets of spatial features whose instances are frequently located together. Mining colocation patterns is important in many applications such as identifying relationships between diseases...
详细信息
Colocation patterns refer to subsets of spatial features whose instances are frequently located together. Mining colocation patterns is important in many applications such as identifying relationships between diseases and environmental factors, but is computationally challenging due to the large number of instances and candidate patterns. Existing algorithms are mostly sequential, and thus can be insufficient for big spatial event data. Recently, parallel colocation mining algorithms have been developed based on the Map-reduce framework, which is economically expensive. Another work proposed a gpu algorithm based on iCPI tree, but assumes that the number of neighbors for each instance is within a small constant, and thus cannot be used when instances are dense and unevenly distributed. To address these limitations, we recently proposed grid-based gpu colocation mining algorithms that include a novel cell-aggregate-based upper bound filter, and two refinement algorithms. In this paper, we provide theoretical analysis of running time. Furthermore using gpu profiling, we identify our recent gpu implementation, gpu-grid-join, as a memory bound problem and to address its bottlenecks, we proposes gpu-grid-join+, an optimized gpu algorithm. Our experimental results on real world data shows that gpu-grid-join+ achieves 4 to 12-fold speedup over gpu-grid-join both running on Nvidia P100 gpu as well as 56 to 126-fold speedup over OpenMP implementation over Intel(R) Xeon(R) CPU with 12 cores. Also for synthetic data, the speedup is in ranges 3 to 7-fold and 9 to 42-fold respectively.
We consider the problem of real-time gpu rendering of algebraic surfaces defined by Bezier tetrahedra. These surfaces are rendered directly in terms of their polynomial representations, as opposed to a collection of a...
详细信息
We consider the problem of real-time gpu rendering of algebraic surfaces defined by Bezier tetrahedra. These surfaces are rendered directly in terms of their polynomial representations, as opposed to a collection of approximating triangles, thereby eliminating tessellation artifacts and reducing memory usage. A key step in such algorithms is the computation of univariate polynomial coefficients at each pixel;real roots of this polynomial correspond to possibly visible points on the surface. Our approach leverages the strengths of gpu computation and is highly efficient. Furthermore, we compute these coefficients in Bernstein form to maximize the stability of root finding, and to provide shader instances with an early exit test based on the sign of these coefficients. Solving for roots is done using analytic techniques that map well to a SIMD architecture, but limits us to fourth order algebraic surfaces. The general framework could be extended to higher order with numerical root finding.
暂无评论