R. Strzodka. Virtual 16 bit precise operations on RGBA8 textures. In Proceedings VMV'02, 2002. Home/Search Document Details and Download
Summary Related Articles Check |
This paper is cited in the following contexts:
Accurate Dense Stereo Reconstruction Using Graphics Hardware - Zach, Klaus, Hadwiger.. (2003) (1 citation) (Correct)
....implementations. They concluded that an efficient memory interface (especially when transfering data from graphics memory into main memory) is still an unsolved issue. For the same reason our implementation is designed to minimize memory traffic between graphics hardware and main memory. Strzodka[18] discusses the emulation of higher precision numbers with several 8 bit color channels. We faced a similar problem of manipulating large integer values and storing them in the frame buffer and texture maps for later use. 3 Overview of Our Method The input for our procedure are two relatively ....
R. Strzodka. Virtual 16 bit precise operations on rgba8 textures. In Proc. of Vision, Modeling and Visualization 2002.
Quality Issues of Hardware-Accelerated High-Quality.. - Hadwiger, Hauser, Möller (2003) (Correct)
....hierarchical way that is also able to achieve higher precision results. However, it requires to know the numerical subrange that is actually used by a given set of rendering passes. Another approach that can be used to improve precision on lower precision hardware has also been suggested recently [14]. 3.1 Preserving internal precision across passes Provided that internal computations are done by the hardware in higher precision than the external precision of the color buffer, a multi pass approach can be used in order to preserve this precision across passes to a certain extent, even if ....
R. Strzodka. Virtual 16 bit precise operations on rgba8 textures. In Proceedings of Vision, Modeling, and Visualization 2002.
In International Conference on Image Analysis and.. - Graphics Hardware.. Self-citation (Strzodka) (Correct)
No context found.
R. Strzodka. Virtual 16 bit precise operations on RGBA8 textures. In Proceedings VMV'02, 2002.
A Graphics Hardware Implementation of the Generalized.. - Strzodka, Ihrke, Magnor (2003) Self-citation (Strzodka) (Correct)
No context found.
R. Strzodka. Virtual 16 bit precise operations on RGBA8 textures. In Proceedings VMV'02, 2002.
A Graphics Hardware Implementation of the Generalized.. - Strzodka, Ihrke, Magnor Self-citation (Strzodka) (Correct)
....sR: t uR . On future graphics hardware with 16 bit and 32 bit number formats, this limitations will become obsolete. Currently, in order to be able to use large outlines and fine weight gradation, we have implemented an unsigned version of the virtual 16 bit fixed point format for RGBA8 textures [27], which emulates 16 bit precise operations on DX8 graphics hardware. This costs a factor of about two due to the doubled amount of data to be transfered, and an additional factor of two due to the need of re transferring each intermediate sum to texture memory, since the 16 bit emulation cannot be ....
....and the pose index pose are stored as 16 bit fixedpoint numbers represented by two 8 bit color channels each. In this way we may index up to o pCpL different poses. The update (Eq. 4) itself can be implemented in a way very similar to the subtraction in the virtual 16 bit format [27]. After all poses have been processed, the texture F contains at each the index of the most probable pose for this location. This final result is transferred to main memory, where the largest values and pose indices are extracted from . Here we see the great advantage in the use of ....
R. Strzodka. Virtual 16 bit precise operations on RGBA8 textures. In Proceedings VMV'02, 2002.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC