Interpolated Cumulative Distribution Functions for Image Contrast

Solution Overview

Problem

Current digital photographic systems using histogram equalization and adaptive histogram equalization techniques suffer from computational intensity and produce visible artifacts along boundaries due to bilinear interpolation, degrading image quality when adjacent regions have different cumulative distribution functions.

Innovation Solution

A method is provided to compute and interpolate cumulative distribution functions (CDFs) for pixel regions, creating an interpolated CDF that eliminates artifacts by representing CDFs with curve-fit coefficients and interpolating parameters between surrounding regions, thereby improving image quality and reducing computational effort.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If bilinear interpolation is used to compute effective CDF for pixels, then computational effort is reduced, but visible artifacts appear along boundaries of fixed regions

Engineering Contradiction:

Improvecomputational effortVSAvoidvisible artifacts

Core Design Contradiction:

ProductivityVSObject-affected harmful factors

Solution Approach 1:

The image is divided into multiple fixed regions, and a unique CDF is computed for each region. This segmentation allows the use of pre-computed CDFs for interpolation while maintaining regional accuracy, reducing computational effort without sacrificing quality.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Each fixed region has its own locally optimized CDF that reflects the specific characteristics of that region. This local quality approach ensures that interpolation between regions maintains accuracy while reducing the need to compute unique CDFs for every pixel.

Inventive Principle:

Principle #3Local quality

2Object-affected harmful factors

If a unique CDF is computed for each pixel region, then image quality is improved, but computational intensity increases

Engineering Contradiction:

Improveimage qualityVSAvoidcomputational intensity

Core Design Contradiction:

Object-affected harmful factorsVSProductivity

Solution Approach 1:

The image is divided into multiple fixed regions, and a unique CDF is computed for each region. This segmentation allows the use of pre-computed CDFs for interpolation while maintaining regional accuracy, reducing computational effort without sacrificing quality.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

CDFs are pre-computed for each fixed region before processing individual pixels. This preliminary action allows subsequent pixel processing to use interpolation of pre-computed values rather than computing unique CDFs for each pixel, significantly reducing computational intensity.

Inventive Principle:

Principle #10Preliminary action

3Object-affected harmful factors

If adaptive histogram equalization is implemented, then perceived image detail and quality are improved, but computational complexity increases significantly

Engineering Contradiction:

Improveperceived image detailVSAvoidcomputational complexity

Core Design Contradiction:

Object-affected harmful factorsVSDevice complexity

Solution Approach 1:

The image is divided into multiple fixed regions, and a unique CDF is computed for each region. This segmentation allows the use of pre-computed CDFs for interpolation while maintaining regional accuracy, reducing computational effort without sacrificing quality.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

CDFs are pre-computed for each fixed region before processing individual pixels. This preliminary action allows subsequent pixel processing to use interpolation of pre-computed values rather than computing unique CDFs for each pixel, significantly reducing computational intensity.

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS11620754B2System, method, and computer program for adjusting image contrast using parameterized cumulative distribution functions

Publication Date: 2023.04.04 DUELIGHT LLC

AI summary

A system and method are provided for optimizing histogram cumulative distribution function curves. In use, a first cumulative distribution function for a first histogram of a first pixel region of a first image is computed, and a first set of parameters for the first cumulative distribution function is extracted. A second cumulative distribution function for a second histogram of a second pixel region of the first image is computed, and a second set of parameters for the second cumulative distribution function is extracted. An interpolated cumulative distribution function comprising interpolated parameters calculated by interpolating between the first set of parameters and the second set of parameters is created. Additionally, a first equalized pixel in a second image based on a first input pixel in the first image and the interpolated cumulative distribution function is generated.