Spectral Color Transformation Using Hyperspectral Reference Atoms
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional electronic image sensors are inherently monochromatic, losing chrominance information and requiring color filter arrays (CFAs) to approximate color detection, which leads to inefficiencies and noise in color conversion, especially when using non-RGB imaging sensors like CYM.
Innovation Solution
A system and method for constructing a spectral reference dataset of naturally-occurring hyperspectral color samples, termed 'spectral atoms,' to transform image data between color spaces, such as CYM to RGB, preserving light collection efficiency and signal-to-noise ratio while accurately converting to a target color space.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If color information is captured using conventional RGB or CMYK color spaces, then the data can be stored and processed efficiently, but the spectral information is lost and color accuracy under different lighting conditions deteriorates
Solution Approach 1:
The patent transitions from conventional 3-dimensional color spaces (RGB) or 4-dimensional color spaces (CMYK) to a 10-dimensional Lab color space. This dimensional expansion allows simultaneous representation of both color information and spectral characteristics, resolving the contradiction between efficient data storage and preservation of spectral information for accurate color measurement under varying lighting conditions.
2Measurement precision
If spectral data is fully captured and stored, then complete color information is preserved, but data processing complexity and storage requirements increase significantly
Solution Approach 1:
The patent extracts and separates spectral information from full spectral data by using a spectral estimator that derives spectral characteristics from the 10-dimensional Lab color space data. This extraction approach preserves essential spectral information for accurate color measurement while eliminating the need to store and process complete spectral curves, thereby reducing data processing complexity and storage requirements.
3Productivity
If conventional color spaces are used for image data, then storage and transmission are efficient, but color constancy under varying illumination conditions deteriorates
Solution Approach 1:
The patent performs preliminary action by capturing and storing illuminant information (light source characteristics) alongside the 10-dimensional Lab color space data during the image acquisition stage. This preliminary capture of illumination conditions enables subsequent spectral estimation and color correction processes to compensate for varying lighting conditions, maintaining color constancy while preserving storage efficiency through the compact 10-dimensional representation.
Data Source
Figure 1
Figure 2
Figure 3A
AI summary
A system comprising at least one hardware processor; and a non-transitory computer readable storage medium having stored thereon program instructions, the program instructions executable by the at least one hardware processor to: receive a set of spectral datapoints obtained from a plurality of images of natural scenes, wherein the images are captured using a hyperspectral imaging device, perform data sampling to obtain a subset of the set of spectral datapoints that is representative of naturally-occurring spectral samples, and said subset of datapoints over (i) a known spectral response profile of a source color space, and (ii) a known spectral response profile of a target color space, to obtain corresponding datasets of source spectral reference atoms and target spectral reference atoms.