Log Color Space Illumination Invariance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing chromaticity representations in image analysis are not fully illumination invariant, particularly when dealing with bi-illuminant scenarios, leading to inaccuracies in material identification and shadow removal across varying illumination conditions.
Innovation Solution
The development of a bi-illuminant dichromatic reflection model (BIDR) in a log color space, which uses a spectral ratio to represent color changes from full shadow to full illumination, allowing for the projection of pixels onto a chromaticity plane that is aligned to be illumination invariant, utilizing techniques such as entropy minimization for optimal alignment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional chromaticity representation (r, g values) is used, then the representation is simple and easy to compute, but it is not illumination invariant and cannot accurately represent material properties under varying illumination conditions
Solution Approach 1:
The patent transforms the traditional linear chromaticity representation into a logarithmic chromaticity space, changing the mathematical parameters from linear ratios to logarithmic transformations. This parameter change enables illumination invariance because the log transform converts multiplicative illumination effects into additive components that can be separated and removed, while maintaining computational feasibility through standardized image processing operations.
Solution Approach 2:
The patent introduces an intermediate chromaticity plane as a mediator between the original RGB color space and the final illumination-invariant material representation. This intermediate plane serves as a transformation space where illumination and reflectance components can be separated through geometric projection, acting as a bridge that enables accurate material identification without requiring direct complex calculations in the original color space.
2Reliability
If log transform of chromaticity values is applied, then illumination invariance is improved for Planckian illuminants, but the representation is still limited and only approximates truly illuminant invariant representation
Solution Approach 1:
The patent implements a dynamic chromaticity plane orientation that adapts to different illumination conditions, particularly bi-illuminant scenarios. Instead of using a fixed chromaticity plane, the system dynamically determines the optimal plane orientation based on the specific illumination environment, allowing accurate material representation whether under single Planckian illuminants or complex multi-illuminant conditions like shadow and highlight regions.
Solution Approach 2:
The patent creates a universal chromaticity representation framework that works across multiple illumination types and scenarios. The method is designed to handle both traditional Planckian illuminants and complex bi-illuminant situations, making it universally applicable to diverse real-world lighting conditions while maintaining a unified mathematical framework rather than requiring separate processing for different illumination types.
3Ease of operation
If chromaticity plane projection is used to remove shadows, then shadow removal capability is improved, but material identification accuracy deteriorates under complex illumination conditions
Solution Approach 1:
The patent performs preliminary separation of illumination and reflectance components in the chromaticity space before final material identification. By pre-processing the image data to isolate and remove illumination effects (including shadows and highlights) in the intermediate chromaticity plane, the system prepares clean, illumination-invariant material data for subsequent identification, ensuring both effective shadow removal and accurate material characterization.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
In a first exemplary embodiment of the present invention, an automated, computerized method for generating an illumination invariant, chromaticity representation of an image is provided. The method of the present invention comprises the steps of providing an image comprising an array of pixels, each pixel having N band color values, transforming the N bands to log color space values in a log color space, generating a bi-illuminant chromaticity plane in the log color space and projecting the log color space values to the chromaticity the plane to provide chromaticity representation values corresponding to the pixels of the image.