Clustered principal components for precomputed radiance transfer
US7167176B2 · kind B2 · utility
Assignee
Inventors
Key dates
| Filing date | Aug 15, 2003 |
| Grant date | Jan 23, 2007 |
| Priority date | — |
| Expiry date | Jun 3, 2024 |
Classification
- Technology area (CPC G)Physics
- CPC primaryG06T15/50
- WIPO fieldComputer technology
- WIPO sectorElectrical engineering
Abstract
Real-time processing includes per-point transfer matrices forming a high-dimensional surface signal that is compressed using clustered principal component analysis (CPCA). CPCA partitions multiple samples into fewer clusters, each cluster approximating the signal as an affine subspace. Further, source radiance is input to a processor, which approximates source radiance using spherical harmonic basis to produce a set of source radiance coefficients. A graphics processing unit (GPU) processes the source radiance coefficients through the transfer matrix model for each cluster. The result of such processing is the exit radiance, which parameterizes the radiance leaving the surface of the object at each point, thus producing the shading for each point of the virtual object in real time.
Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.