Reduced Lookup Table for Gamma Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing systems face challenges with nonlinear responses in display and sensor systems, leading to issues like loss of definition in low-light areas, and the computational expense of gamma correction, particularly due to the large size of lookup tables required for high-bit pixel values.
Innovation Solution
A method and apparatus that construct a reduced-size lookup table by partitioning the input range into sections, using a set of boundary values and index offset values to map input values to output values with fewer entries, thereby reducing storage and computational requirements while maintaining image quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a full-resolution lookup table is used for high-bit pixel values (e.g., 14-bit), then mapping precision is maintained, but storage requirements and loading time become prohibitively large
Solution Approach 1:
The input pixel value range is divided into multiple sections based on the slope characteristics of the gamma curve. Each section is mapped to a separate sub-table with appropriate resolution, allowing precise mapping where needed while reducing overall storage requirements through selective downsampling in less critical regions.
Solution Approach 2:
Different sections of the gamma curve are assigned different mapping resolutions based on their local characteristics. Regions with steep slopes (requiring higher precision) use higher-resolution sub-tables, while regions with gentle slopes use lower-resolution sub-tables, optimizing the balance between image quality and storage efficiency.
2Measurement precision
If a full-resolution lookup table is used, then mapping precision is maintained, but loading time becomes too long for run-time applications
Solution Approach 1:
The large lookup table is segmented into multiple smaller sub-tables corresponding to different sections of the gamma curve. This allows the system to load only the necessary sub-tables for the current image processing task rather than loading the entire table, significantly reducing loading time while maintaining precision where required.
Solution Approach 2:
The gamma curve is pre-analyzed to identify critical sections that require high-resolution mapping. Only these critical sections are included in the final lookup table structure, eliminating unnecessary data and reducing loading time while preserving mapping precision in the most important regions.
3Quantity of substance
If the lookup table resolution is reduced to decrease storage requirements, then storage and loading time are improved, but mapping precision deteriorates
Solution Approach 1:
The lookup table uses variable resolution across different sections, with higher precision allocated to regions of the gamma curve where the slope is steep and precision is most critical for image quality, and lower precision in regions where the slope is gentle and precision requirements are lower.
Solution Approach 2:
The system dynamically selects which sub-tables to load and use based on the characteristics of the input image and the specific gamma curve being applied, adapting the resolution and storage usage to the actual processing needs rather than using a fixed high-resolution table for all cases.
Data Source
AI summary
A method of processing an image includes selecting, for each of a plurality of picture element values in at least a portion of the image, one among a plurality of offset values. For each of the plurality of picture element values, an index value is obtained based on (A) the selected offset value and (B) a portion of the picture element value. For each of the plurality of picture element values, an entry is retrieved from a lookup table according to the corresponding index value.