Neural Color Coordinate Estimation from Normalized Spectral Responses
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for obtaining color coordinates of light sources using RGB values and conversion matrices result in deviations due to variations in reflectivity and intensity, leading to inaccurate color coordinate estimation.
Innovation Solution
A light source color coordinate estimation system utilizing a neural network with photo detectors, a normalization calculation circuit, and a neural network that normalizes energy integral values and adjusts weights through a deep learning process to accurately estimate color coordinates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If RGB values are converted to color coordinates using a conversion matrix, then color coordinate calculation is achieved, but measurement precision deteriorates due to deviations caused by reflectivity and intensity variations
Solution Approach 1:
The patent changes the input parameters from RGB values to spectral power distribution (SPD) data obtained from multiple photo detectors with different detection wavebands. This parameter transformation allows the system to capture reflectivity and intensity characteristics more accurately, thereby improving color coordinate estimation precision and reducing deviations
Solution Approach 2:
The patent replaces the traditional matrix conversion method with a neural network-based estimation system. The neural network processes normalized energy integral values from multiple photo detectors to directly output color coordinates, substituting the mechanical matrix multiplication process with an intelligent estimation system that adapts to various light source characteristics
2Measurement precision
If multiple photo detectors with different detection wavebands are used, then measurement precision improves, but device complexity increases
Solution Approach 1:
The patent employs a universal neural network model that can process input from multiple photo detectors with different detection wavebands. The neural network serves multiple functions: it normalizes the energy integral values from different detectors, processes the spectral data, and outputs color coordinates. This multi-functionality justifies the use of multiple photo detectors while managing system complexity through a unified processing architecture
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The system achieves precise color coordinate estimation by normalizing energy integral values and adjusting neural network weights, reducing deviations caused by beam intensity and surface reflectivity, thereby enhancing accuracy.
Implementation Method 1
The photo detectors generate spectral responses after receiving a beam emitted by a light source. The spectral responses respectively have different detection wavebands and energy integral values corresponding to the detection wavebands.
Data Source
AI summary
A light source color coordinate estimation system and a deep learning method thereof are provided. The light source color coordinate estimation system includes a plurality of photo detectors, a normalization calculation circuit, and a neural network. The photo detectors generate spectral responses after receiving a beam emitted by a light source. The spectral responses respectively have different detection wavebands and energy integral values corresponding to the detection wavebands. The normalization calculation circuit respectively divides the energy integral values by a largest of the energy integral values to generate a plurality of normalized energy integral values. An input end of the neural network receives the normalized energy integral values and converts the normalized energy integral values into an estimated color coordinate. An output end of the neural network outputs the estimated color coordinate.