Local Contrast Value Determination for Digital Image Sharpness

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for determining image sharpness using contrast values are flawed as they can be skewed by high contrast due to luminance differences, image sensor noise, and optical effects, making it difficult to adjust camera settings effectively.

Innovation Solution

A method that calculates a local contrast value based on edge occurrence and noise estimation using edge detection algorithms, which filters out luminance differences and noise, providing a noise-insensitive measure for adjusting camera settings.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional contrast value methods are used to determine image sharpness, then the calculation is simple based on pixel values, but the result is skewed by luminance differences, sensor noise, and optical effects

Engineering Contradiction:
Improvesharpness measurement accuracyVSAvoidcontrast calculation complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent extracts only the edge-related pixel value differences from the complete image data, ignoring luminance differences in uniform areas. By applying edge detection algorithms, the method isolates regions with actual structural information where contrast is meaningful for sharpness assessment, thereby eliminating the skewing effects of luminance variations and sensor noise in non-edge regions

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent introduces edge detection algorithms as an intermediary step between raw pixel data and contrast calculation. This intermediary process filters out harmful factors (luminance differences, noise) before the contrast value is computed, ensuring that only edge-related contrast information contributes to the sharpness measurement

Inventive Principle:
Principle #24Intermediary (Mediator)

2Measurement precision

If edge detection algorithms are applied to filter out luminance differences and noise, then the sharpness measurement becomes more accurate, but the computational complexity increases

Engineering Contradiction:
Improvesharpness measurement accuracyVSAvoidcomputational energy consumption
Core Design Contradiction:
Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The patent segments the image processing task into distinct stages: edge detection followed by contrast calculation on edge regions only. This segmentation allows the computationally intensive edge detection to be performed once, and then the simplified contrast calculation to be applied only to edge-related pixel values, reducing redundant computations compared to calculating contrast over the entire image

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies contrast calculation selectively only to regions identified as edges, rather than performing the calculation on all pixel pairs in the image. This partial action approach concentrates computational resources on the most relevant regions for sharpness assessment, reducing overall computational energy consumption while maintaining measurement accuracy

Inventive Principle:
Principle #16Partial or excessive action

3Area of stationary object

If contrast is calculated based on all pixel values including uniform areas, then the calculation covers the entire image, but sensor noise creates false contrast values

Engineering Contradiction:
Improveimage coverage areaVSAvoidcontrast value reliability
Core Design Contradiction:
Area of stationary objectVSMeasurement precision

Solution Approach 1:

The patent extracts and isolates edge-related pixel value differences from the complete image data, deliberately excluding uniform areas where sensor noise dominates. By focusing calculation only on regions with detected edges, the method ensures that contrast values reflect actual image features rather than noise artifacts, thereby improving reliability while maintaining comprehensive image coverage through the edge detection process

Inventive Principle:
Principle #2Taking out (Extraction)

4Ease of operation

If global contrast is used to assess sharpness, then the calculation is straightforward, but high contrast in bright areas does not indicate actual sharpness

Engineering Contradiction:
Improvesharpness assessment simplicityVSAvoidsharpness measurement accuracy
Core Design Contradiction:
Ease of operationVSMeasurement precision

Solution Approach 1:

The patent introduces edge detection algorithms as an intermediary step between raw pixel data and contrast calculation. This intermediary process filters out harmful factors (luminance differences, noise) before the contrast value is computed, ensuring that only edge-related contrast information contributes to the sharpness measurement

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent applies contrast calculation selectively to local edge regions rather than globally across the entire image. By making the contrast measurement local to edge structures, the method captures meaningful sharpness information in specific regions while ignoring misleading high-contrast areas such as bright skies or uniform regions that do not reflect actual focus quality

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP3503528B1Determination of a contrast value for a digital image
Publication Date: 2020.10.07 AXIS
  • EP3503528B1 patent drawingFigure 1
  • EP3503528B1 patent drawingFigure 2
  • EP3503528B1 patent drawingFigure 3

AI summary

A method for determining a local contrast value for a digital image captured by an image sensor in a camera is presented. The method comprising: applying an edge detection algorithm to image data of the digital image, thereby obtaining a data set pertaining to edges in the digital image; calculating, based on said data set, an edge occurrence value for the digital image; estimating, based on image data of the digital image and a noise model of the image sensor, an estimated image sensor noise for the digital image; and computing the local contrast value as a relationship between the edge occurrence value of the digital image and the expected edge occurrence value for the digital image. Also a method for adjusting a focus setting of a camera using the local contrast value for images captured by the camera is presented.