GPU-Based Optimization of a Free-Viewpoint Video System

• Autores: Neal Orman, Hansung Kim, Ryuuki Sakamoto, Tomoji Toriyama, Kiyoshi Kogure, Robert Lindeman
• Localización: ELCVIA. Electronic letters on computer vision and image analysis, ISSN-e 1577-5097, Vol. 7, Nº. Extra 2, 2008 (Ejemplar dedicado a: Special Issue on Computational Modelling of Objects Represented in Images), págs. 120-133
• Idioma: inglés
• Enlaces
  • Texto completo  1    2  
• Resumen

  • We present a method for optimizing thereconstruction and rendering of 3D objects from multiple imagesby utilizing the latest features of consumer-level graphicshardware based on shader model 4.0. We accelerate visual hullreconstruction by rewriting a shape-from-silhouette algorithm to execute on the GPU's parallel architecture. Rendering isoptimized through the application of geometry shaders togenerate billboarding microfacets textured with captured images.We also present a method for handling occlusion in the cameraselection process that is optimized for execution on the GPU.Execution time is further improved by rendering intermediate results directly to texture to minimize the number of data transfers between graphics and main memory. We show our GPUbased system to be significantly more efficient than a purely CPUbasedapproach, due to the parallel nature of the GPU, while maintaining graphical quality.