Illuminant-Invariant Model Estimation via Normalized Radiance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing electronic devices that perform computer vision, machine learning, and image processing tasks are vulnerable to changes in illuminance, leading to inconsistencies in RGB image sequences, which affects accuracy in applications like 3D reconstruction, feature matching, and object detection.
Innovation Solution
The method involves determining normalized radiance and reliability images based on a camera response function, extracting features, and optimizing models using these features and reliability images to achieve illuminant-invariant model estimation, which is robust to changes in the illuminant environment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional RGB image processing is used, then the processing is simple and fast, but the accuracy deteriorates under illuminance changes
Solution Approach 1:
The patent introduces normalized radiance as an intermediary representation that mediates between the raw RGB image data and the final processing tasks. By transforming RGB values through the camera response function to obtain normalized radiance, the system creates a representation that is invariant to illuminance changes, thereby improving accuracy without requiring fundamental changes to the overall processing pipeline
Solution Approach 2:
The patent changes the parameter space by transforming from RGB color space to normalized radiance space using the camera response function. This parameter transformation allows the processing to be performed in a space where illuminance variations do not affect the measurements, thereby improving accuracy while maintaining similar computational complexity
2Reliability
If illuminant-invariant processing is implemented, then robustness to illuminance changes is improved, but computational complexity increases
Solution Approach 1:
The patent applies preliminary action by computing the camera response function and transforming images to normalized radiance space before performing the actual processing tasks. By pre-processing the images to remove illuminance dependencies, the subsequent processing steps can proceed with standard algorithms, thereby achieving robustness without proportionally increasing overall computational complexity
Solution Approach 2:
The normalized radiance serves as an intermediary that enables robust processing. By introducing this intermediate representation that is specifically designed to be invariant to illuminance changes, the system achieves reliability improvement while the computational overhead is confined to the transformation step rather than affecting all subsequent operations
3Measurement precision
If features are extracted from normalized radiance, then accuracy under varying illuminance is improved, but processing time increases
Solution Approach 1:
The patent changes the parameter space to normalized radiance for feature extraction, which improves accuracy by removing illuminance variability from the features. The processing time increase is mitigated by the fact that the transformation to normalized radiance can be performed efficiently using the pre-determined camera response function, and the improved accuracy reduces the need for iterative refinement in subsequent matching steps
Data Source
AI summary
A method is described. The method includes determining normalized radiance of an image sequence based on a camera response function (CRF). The method also includes determining one or more reliability images of the image sequence based on a reliability function corresponding to the CRF. The method further includes extracting features based on the normalized radiance of the image sequence. The method additionally includes optimizing a model based on the extracted features and the reliability images.