Adaptive Image Sharpening Using Pixel-Wise Contrast and Brightness
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image sharpening techniques fail to dynamically adjust sharpening strength based on the specific content of the image, leading to suboptimal text readability and overall image quality, particularly in scenarios where text is converted or captured as an image.
Innovation Solution
Adaptive image sharpening techniques determine sharpening strengths for individual pixels based on local contrast and brightness parameters, using a 2D look-up table (LUT) in the display engine hardware, which is trained offline to provide tailored sharpening for specific use-cases like gaming or video calls, without real-time AI model execution.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If existing image sharpening techniques are used, then image sharpening is applied, but text readability and visual quality remain suboptimal due to lack of dynamic adjustment
Solution Approach 1:
The patent applies different sharpening strengths to different regions of the image based on local contrast and brightness characteristics. Text regions with high contrast receive stronger sharpening, while low-contrast regions receive weaker sharpening, achieving optimal text readability without introducing artifacts in other areas.
Solution Approach 2:
The sharpening strength is dynamically adjusted for each pixel based on its local contrast and brightness parameters. The system computes these parameters in real-time and uses them to determine the appropriate sharpening strength, enabling adaptive response to varying image content rather than applying a fixed sharpening strength.
2Manufacturing precision
If real-time AI model execution is used for sharpening, then tailored sharpening for specific use-cases is achieved, but computational overhead increases
Solution Approach 1:
The AI model is trained offline in advance to learn the relationship between local contrast/brightness parameters and optimal sharpening strengths. The trained model generates lookup tables that map these parameters to sharpening strengths, eliminating the need for real-time AI model execution while maintaining tailored sharpening quality for specific use-cases.
Solution Approach 2:
Instead of executing the complex AI model in real-time, the system uses a simplified lookup table that copies the essential information from the trained model. The lookup table stores pre-computed sharpening strengths for different local contrast and brightness combinations, allowing fast retrieval without the computational burden of real-time model execution.
3Productivity
If fixed sharpening strength is applied, then processing is simple, but image quality deteriorates due to inability to adapt to different content
Solution Approach 1:
The image is segmented into individual pixels or small regions, and each segment is processed independently based on its local characteristics. The system computes local contrast and brightness for each pixel and uses this information to determine the appropriate sharpening strength, allowing simple processing at the pixel level while achieving high overall image quality through adaptive sharpening.
Data Source
AI summary
Systems, apparatus, articles of manufacture, and methods to implement adaptive image sharpening are disclosed. An example apparatus disclosed herein computes respective brightness parameters and respective contrast parameters for corresponding pixels of a first image. The disclosed example apparatus also determines respective sharpening strength parameters for the corresponding pixels based on the respective brightness parameters and the respective contrast parameters. The disclosed example apparatus further generate a second image based on the first image and the sharpening strength parameters.