Scene Adaptive Global Brightness Contrast Using Gamma-Corrected Histograms
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing Global Brightness Contrast Enhancement (GBCE) algorithms fail to optimally adjust brightness and contrast based on image content, leading to noise amplification in dark images, color desaturation, over-saturation in highlight regions, and poor generalization across various scene conditions, while traditional histogram-based methods result in suboptimal image quality.
Innovation Solution
A method that uses Gamma-corrected histograms and ancillary data to determine gain control, computes scene-adaptive tone points, blends with histogram equalization information, and generates a look-up table for enhancing global brightness contrast, incorporating scene-specific gamma curves and tone mapping to optimize brightness and contrast.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Illumination intensity
If traditional histogram-based GBCE algorithms are used to improve brightness and contrast, then brightness and contrast enhancement is achieved, but noise is amplified in dark/low light images and color becomes de-saturated
Solution Approach 1:
The patent applies different tone mapping strategies to different regions of the image based on their brightness characteristics. Shadow regions receive different processing than mid-tone and highlight regions, allowing optimal enhancement in each region while preserving natural appearance and avoiding noise amplification in dark areas.
Solution Approach 2:
The system dynamically adjusts tone mapping parameters based on scene conditions and histogram analysis. By changing the tone curve parameters adaptively, the system optimizes brightness enhancement while controlling noise amplification and color saturation across different imaging conditions.
2Illumination intensity
If histogram equalization is applied to improve brightness and contrast, then brightness and contrast enhancement is achieved, but the image looks washed out and content information is lost
Solution Approach 1:
Instead of applying uniform histogram equalization that flattens the distribution, the patent inverts the approach by preserving natural histogram characteristics while applying selective tone mapping. This maintains the natural distribution of pixel values and preserves content information while still improving visual appearance.
Solution Approach 2:
The system applies different tone mapping characteristics to different brightness regions, preserving the natural content distribution in each region rather than uniformizing the entire histogram. This maintains scene information while improving overall visual quality.
3Illumination intensity
If global brightness enhancement is applied to underexposed images, then brightness improvement is achieved, but highlight regions become over-saturated and mid-tones are compromised
Solution Approach 1:
The patent segments the tone range into shadow, mid-tone, and highlight regions, applying different enhancement strategies to each. This allows brightness improvement in underexposed areas while protecting highlight regions from over-saturation and preserving mid-tone accuracy.
Solution Approach 2:
The system dynamically adjusts tone mapping parameters based on the specific brightness distribution of each region. By changing parameters locally for different tone ranges, the system optimizes enhancement in underexposed areas without compromising highlight or mid-tone quality.
4Adaptability or versatility
If adaptive tone mapping is used to preserve content in bright and dark regions, then dynamic range preservation is achieved, but the algorithm complexity increases
Solution Approach 1:
The system pre-computes tone mapping curves for different scene conditions based on training data. During actual processing, these pre-computed curves are selected and applied, avoiding the need for complex real-time optimization while maintaining scene adaptiveness.
Solution Approach 2:
The patent uses learned tone mapping curves from training data that capture optimal processing for different scene types. These pre-learned curves are copied and applied to new images, providing scene adaptiveness without requiring complex real-time computation.
Data Source
AI summary
A method and apparatus for enhancing at least one of video and image quality using an exposure aware scene adaptive global brightness contrast. The method includes determining if gain control is needed utilizing Gamma corrected histogram and ancillary data, if gain control is not required, computing a set of scene adaptive tone points and blending with information from histogram equalization if needed, if gain control is required, estimating gain value and utilizing the gain value in computing the gain table, and accordingly, generating a look-up table for enhancing global brightness contrast utilizing the gain table, the tone table and the information from histogram equalization.