Digital Image Highlight Restoration via Saturation Map
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing digital image processing methods fail to accurately correct saturated pixels in highlights, leading to incorrect color representation due to signal saturation, which affects both color information and dynamic range.
Innovation Solution
A method using a saturation map to identify and correct saturated pixels by generating replacement values from nearby unsaturated pixels, combining restoration and neutralization techniques, and employing a correction table to address signal saturation issues.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If color desaturation or neutralization is applied to correct saturated pixels, then color accuracy is improved, but dynamic range information is lost
Solution Approach 1:
The image processing is segmented into multiple stages: saturation detection, selective pixel classification (restorable vs. non-restorable), restoration processing for pixels with at least one unsaturated channel, and neutralization for pixels with all saturated channels. This segmentation allows different processing strategies to be applied to different pixel types, preserving dynamic range information where possible while correcting color accuracy where needed.
Solution Approach 2:
Different correction quality levels are applied to different pixels based on their saturation characteristics. Pixels with at least one unsaturated channel receive restoration processing that preserves more information, while only pixels with all three channels saturated receive neutralization. This local differentiation optimizes the balance between color accuracy and dynamic range preservation for each pixel.
2Measurement precision
If restoration processing is applied to pixels with all saturated channels, then color accuracy may be improved, but processing complexity increases
Solution Approach 1:
The pixel population is segmented into two categories: restorable pixels (at least one unsaturated channel) and non-restorable pixels (all channels saturated). This segmentation simplifies the overall processing logic by allowing the majority of pixels to follow a straightforward restoration path, while only a minority require the more complex neutralization process.
Solution Approach 2:
The patent applies restoration processing to all pixels with at least one unsaturated channel, which is a partial action approach. This avoids the excessive complexity of attempting to restore all saturated pixels uniformly, while still achieving good color accuracy for the majority of affected pixels. The neutralization process is applied only when necessary.
3Reliability
If neutralization is applied to all saturated pixels, then color artifacts are reduced, but color information is lost
Solution Approach 1:
Neutralization is applied locally only to pixels where it is truly necessary (those with all three color channels saturated), rather than universally to all saturated pixels. This localized application minimizes color information loss while still achieving artifact reduction in the critical regions where restoration is impossible.
Solution Approach 2:
The processing pipeline segments pixels into restorable and non-restorable groups, applying different correction strategies. This segmentation ensures that neutralization is reserved for cases where restoration cannot work (all channels saturated), thereby preserving color information in pixels where restoration can maintain both accuracy and information content.
Data Source
AI summary
A method for performing highlight restoration on a digital image includes comparing the pixels in the image with a saturation level value to identify saturated pixels. A saturation map of saturated pixels is generated. Each selected saturated pixel is identified as a restorable pixel only if at least one color channel of the pixel is unsaturated. For each restorable pixel, a group of the closest unsaturated pixels above, below, to the left, and to the right of the select saturated pixel is identified. A replacement pixel value is generated for each saturated color channel of the restorable pixel, using a combination of the pixel values of the unsaturated color channels of the restorable pixel and the pixel values of the corresponding color channels of the nearby unsaturated pixels.