Estamos realizando la búsqueda. Por favor, espere...

Real-time stereo matching using memory-efficient Belief Propagation for high-definition 3D telepresence systems

Abstract: New generations of telecommunications systems will include high-definition 3D video that provides a telepresence feeling. These systems require high-quality depth maps to be generated in a very short time (very low latency, typically about 40 ms). Classical Belief Propagation algorithms (BP) generate high-quality depth maps but they require huge memory bandwidths that limit low-latency implementations of stereo-vision systems with high-definition images. This paper proposes a real-time (latency inferior to 40 ms) high-definition (1280 × 720) stereo matching algorithm using Belief Propagation with good immersive feeling (80 disparity levels). There are two main contributions. The first is an improved BP algorithm with pixel classification that outperforms classical BP while reducing the number of memory accesses. The second is an adaptive message compression technique with a low performance penalty that greatly reduces the memory traffic. The combination of these techniques outperforms classical BP by about 6.0% while reducing the memory traffic by more than 90%.

 Autoría: Pérez J.M., Sánchez P.,

 Fuente: Pattern Recognition Letters, 2011, 32(16), 2250-2253

 Editorial: Elsevier

 Fecha de publicación: 01/12/2011

 Nº de páginas: 4

 Tipo de publicación: Artículo de Revista

 DOI: 10.1016/j.patrec.2011.06.016

 ISSN: 0167-8655,1872-7344

 Proyecto español: TEC2008- 04107

 Url de la publicación: https://doi.org/10.1016/j.patrec.2011.06.016