Real-time Computer-Generated Integral Image Based on GPU-driven Cross Perspective Rendering Pipeline

Yanxin Guan,Xinzhu Sang,Shujun Xing,Yuanhang Li,Yingying Chen,Peng Wang,Duo Chen,Binbin Yan
DOI: https://doi.org/10.1117/1.oe.60.2.023105
IF: 1.3
2021-01-01
Optical Engineering
Abstract:Abstract. It is insufficiently powerful to compute multi-view images at a high enough frame or data rate to support high quality dynamic three-dimensional displays with the current computer software and hardware. According to the correspondence between lens position and elemental image, a real-time computer-generated content generation method for integral imaging based on cross perspective is presented. A GPU-driven rendering pipeline (GDRP) with geometry instancing and instance culling technique is designed, and the number of triangles processed in the rasterization stage is reduced greatly. Several virtual dynamic scenes including various materials and millions of vertices are used to verify the performance of the GDRP method in the experiment. Experimental results show that the frame rate can reach 60 fps with 3840  ×  2160 resolution.
What problem does this paper attempt to address?