Conference 3 Dec - 6 Dec Exhibition 4 Dec - 6 Dec


    Technical Papers

    01 Full Conference1 - Full Conference One Day



    Vectors and Shaders

    Saturday, 06 December

    14:15 - 16:00

    Jasmine Hall

    Automatic Shader Simplification Using Surface Signal Approximation

    We present a new automatic shader simplification method using surface signal approximation. The fragment shader simplification is formulated as a global simplification problem across multiple shader stages. Three simplification rules are proposed and validated on several different shaders. Results show our approach well the performance, accuracy and memory balance.

    Rui Wang, Zhejiang University
    Xianjing Yang, Zhejiang University
    Yazhen Yuan, Zhejiang University
    Wei Chen, Zhejiang University
    Kavita Bala, Cornell University
    Hujun Bao, Zhejiang University

    Deep Shading Buffers on Commodity GPUs

    An efficient algorithm for decoupled sampling of visibility and shading is presented, which runs at interactive rates on current and near-future GPUs. The method maintains a low pixel shading cost, even at very large blurs. This is a key step towards a real-time pipeline for stochastic rendering.

    Petrik Clarberg, Intel Corporation
    Jacob Munkberg, Intel Corporation

    Whippletree: Task-based Scheduling of Dynamic Workloads on the GPU

    Our new programming and processing model for the GPU allows the definition and scheduling of tasks with strongly varying characteristics. Using our scheduling framework, graphics pipelines and parallel programs can achieve vast speedups over traditional execution models.

    Markus Steinberger, Graz University of Technology
    Michael Kenzel, Graz University of Technology
    Pedro Boechat, Graz University of Technology
    Bernhard Kerbl, Graz University of Technology
    Mark Dokter, Graz University of Technology
    Dieter Schmalstieg, Graz University of Technology

    Massively-Parallel Vector Graphics

    We present a feature-rich real-time vector graphics rendering pipeline for GPUs that is massively parallel at every stage. Our renderer employs wide anti-aliasing filters (e.g. 4x4) and high sampling rates (> 4x4x32). We show renderings of complex vector graphics in state-of-the-art quality and performance.

    Francisco Ganacim, IMPA - Instituo Nacional de Matemática Pura e Aplicada
    Rodolfo S. Lima, IMPA - Instituo Nacional de Matemática Pura e Aplicada
    Luiz Henrique de Figueiredo, IMPA - Instituo Nacional de Matemática Pura e Aplicada
    Diego Nehab, IMPA - Instituo Nacional de Matemática Pura e Aplicada

    Hierarchical Diffusion Curves for Accurate Automatic Image Vectorization

    Our approach automatically and efficiently vectorizes a broad class of rasterized input images, resulting in a hierarchical diffusion curve representation. We trace curves along the local maxima in the image Laplacian and bilaplacian domains. Our results use a sparse set of curves and are easy to edit after-the-fact.

    Guofu Xie, University of Montreal
    Xin Sun, Microsoft Research Asia
    Xin Tong, Microsoft Research Asia
    Derek Nowrouzezahrai, University of Montreal

    Back to Top
    Go Back