Tone Mapping Spatial Frequency Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing tone mapping techniques for high-dynamic range (HDR) images on limited-dynamic range (LDR) displays often result in dynamic range compression, halo artifacts, and time-consuming manual adjustments, especially in motion pictures where variations between sequential images exacerbate these issues.
Innovation Solution
Tone mapping each object in an image or animation frame separately based on its surface, separating illumination values into low and high spatial frequency components and adjusting them independently to eliminate halo effects and preserve gradients, allowing for seamless tone mapping across a sequence of frames without introducing low frequency artifacts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If global tone mapping operators are used to convert HDR images to LDR displays, then the entire image can be mapped using consistent parameters, but this results in dynamic range compression causing saturated or too dark areas that look flat and dull
Solution Approach 1:
The patent segments the tone mapping process by separating illumination values into low spatial frequency components (handled by global operator) and high spatial frequency components (handled by local operator), allowing each to be processed with appropriate parameters to avoid the drawbacks of using either operator alone
2Manufacturing precision
If local tone mapping operators are used to preserve local gradients, then illumination value gain changes can be applied locally, but this generates halo or ringing effects around brightly-colored objects against dark backgrounds
Solution Approach 1:
The patent segments the illumination values into spatial frequency components, applying local operators only to high spatial frequency components while using global operators for low spatial frequency components, thereby preserving local gradients without generating halo artifacts
Solution Approach 2:
The patent applies different tone mapping operators to different spatial frequency components: global operators for low frequency components and local operators for high frequency components, optimizing the approach for each component's characteristics
3Manufacturing precision
If manual adjustments are made to correct visual artifacts after global or local tone mapping, then image quality can be improved, but this becomes an arduous and time-consuming task
Solution Approach 1:
The patent segments the tone mapping process to automatically combine the strengths of global and local operators, producing an output that requires minimal or no manual adjustment, thereby reducing time-consuming manual intervention while maintaining high image quality
4Productivity
If variations between sequential images are processed using conventional tone mapping, then each frame can be tone mapped independently, but this exacerbates halo effects and introduces low frequency artifacts in animation sequences
Solution Approach 1:
The patent segments the tone mapping process into frequency-based components and applies different operators to each component, which eliminates halo effects and low frequency artifacts in animation sequences while maintaining efficient frame-by-frame processing capability
Data Source
AI summary
A method for rendering an image including objects defined by surfaces. A rendering application selects an object in a first image and determines a surface of the object. An initial set of illumination values is calculated and is separated into low and high spatial frequency components associated with the surface of the object. The rendering application independently adjusts the illumination values of the low and high spatial frequency components based lighting information in the first image, and generates a modified set of illumination values by combining the adjusted low and high spatial frequency components. The surface of the object is then rendered using the modified set of illumination values. Advantageously, embodiments of the invention provide techniques for rendering an object without introducing halo effects around the object. Additionally, embodiments of the invention provide for rendering a sequence of frames without introducing fluctuations in the low frequency components from across frame.


