Lookup Table Spherical Harmonics Rendering Dynamic Geometry
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Spherical harmonics (SH) based rendering methods face challenges in dynamic geometry due to high data quantity of visibility information, making real-time computation difficult, and fail to properly consider lighting environments and intensity variations.
Innovation Solution
An image processing apparatus and method that uses a lookup table (LUT) to store relations between occlusion vectors and SH coefficients, allowing for pre-computation and real-time rendering by computing occlusion vectors and SH coefficients using a Z-buffer value, normal vector, and bidirectional reflectance distribution function (BRDF), and updating the LUT with occlusion map information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If spherical harmonics based rendering is applied to dynamic geometry, then rendering capability for dynamic objects is improved, but computational complexity increases due to high data quantity of visibility information
Solution Approach 1:
The patent pre-computes visibility information and stores it in a lookup table before runtime. This allows the system to handle dynamic geometry efficiently during rendering by simply querying pre-computed data rather than computing visibility information in real-time, thus resolving the contradiction between handling dynamic objects and maintaining computational efficiency
Solution Approach 2:
The patent creates a simplified representation of visibility information through occlusion maps and stores pre-computed spherical harmonics coefficients in a lookup table. This copied representation allows fast access during rendering without requiring complex real-time computations, enabling dynamic geometry rendering while reducing computational complexity
2Productivity
If conventional SH based methods are used, then rendering speed is improved, but lighting environment accuracy deteriorates due to improper consideration of lighting environments
Solution Approach 1:
The patent computes occlusion vectors that are specific to each point on the object surface, taking into account local lighting conditions and occlusion relationships. This point-specific approach ensures accurate lighting environment representation while maintaining rendering efficiency through the lookup table structure
Solution Approach 2:
The patent transforms the lighting environment representation by projecting occlusion maps onto spherical harmonics basis functions, creating SH coefficients that accurately encode lighting information. This parameter transformation allows the system to maintain both rendering speed and lighting accuracy by working in the spherical harmonics domain
3Adaptability or versatility
If real-time computation of visibility information is performed, then adaptability to dynamic geometry is improved, but processing time increases due to great data quantity
Solution Approach 1:
The patent performs pre-computation of visibility information and spherical harmonics coefficients before runtime, storing results in a lookup table. During actual rendering of dynamic geometry, the system simply queries this pre-computed data, achieving both adaptability to dynamic objects and real-time performance by avoiding expensive runtime computations
Solution Approach 2:
The patent extracts only the essential visibility information needed for rendering by projecting occlusion maps onto spherical harmonics basis functions. This extraction process creates a compact representation that captures the most important lighting and occlusion characteristics while reducing data quantity and processing requirements
Data Source
AI summary
An image processing apparatus may include a storage unit to store a lookup table (LUT) including information on corresponding relations between an occlusion vector related to at least one point of a 3-dimensional (3D) object and a spherical harmonics (SH) coefficient; and a rendering unit to determine a first SH coefficient corresponding to a first occlusion vector related to a first point of the 3D object using the LUT and to determine a pixel value of the first point using the first SH coefficient.