Image Processing Apparatus Contrast Enhancement via Hierarchical Gain Maps

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing image tone processing techniques, such as those using S-curve conversion characteristics, often excessively alter brightness in low-frequency regions like skies or shady concrete, leading to unnecessary contrast improvement.

Innovation Solution

An image processing apparatus generates a hierarchical image, a first gain map based on this image, and a second gain map using the first gain map and a tone characteristic that cancels out the luminance changes, applying the second gain map to the input image to enhance contrast only in high-frequency regions without affecting low-frequency regions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If an S-curve conversion characteristic is applied to improve contrast, then contrast in the image is improved, but brightness in low-frequency regions is excessively changed

Engineering Contradiction:

ImprovecontrastVSAvoidbrightness

Core Design Contradiction:

Manufacturing precisionVSIllumination intensity

Solution Approach 1:

The image is divided into multiple frequency bands using a hierarchical image structure. The gain map generation is performed separately for different frequency components, allowing contrast enhancement in high-frequency regions while preserving brightness in low-frequency regions.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Different tone characteristics are applied to different frequency regions. High-frequency regions receive contrast enhancement through gain map application, while low-frequency regions maintain their original brightness characteristics, achieving localized quality optimization.

Inventive Principle:

Principle #3Local quality

2Illumination intensity

If a compression tone curve is applied and then correction is performed through reduction and enlargement, then brightness change is reduced, but processing complexity increases

Engineering Contradiction:

ImprovebrightnessVSAvoidprocessing complexity

Core Design Contradiction:

Illumination intensityVSDevice complexity

Solution Approach 1:

The hierarchical image structure and gain maps are generated in advance from the input image. This preliminary processing allows the contrast enhancement to be applied directly without requiring subsequent reduction and enlargement operations, simplifying the overall processing flow.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The traditional mechanical approach of reducing and enlarging images to smooth luminance variations is replaced with a computational approach using hierarchical image processing and gain map multiplication, achieving the same effect with less complexity.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentUS10764468B2Image processing apparatus and image processing method

Publication Date: 2020.09.01 CANON KK

AI summary

An apparatus includes an image generation unit configured to generate, from an input image, a hierarchical image including a plurality of images different in frequency band from one another, a first gain map generation unit configured to generate a first gain map with use of the hierarchical image and a first tone characteristic, a second gain map generation unit configured to generate a second gain map with use of the first gain map and a second tone characteristic, and a processing unit configured to apply the second gain map to the input image.