Digital Image Sharpening with Noise Amplification Avoidance

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Traditional digital image sharpening methods amplify noise, leading to reduced image quality and usability, as they often suppress details and introduce blur effects.

Innovation Solution

A method that creates a smoothed version of the original image, calculates a residual image, and applies sharpening techniques to this smoothed image without amplifying noise, allowing for controlled edge enhancement and detail restoration by combining the residual and sharpened images.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If traditional image sharpening methods are applied to enhance edge sharpness, then local contrast between pixels is improved, but noise is significantly amplified

Engineering Contradiction:

Improveedge sharpnessVSAvoidnoise amplification

Core Design Contradiction:

Manufacturing precisionVSObject-affected harmful factors

Solution Approach 1:

The image processing is segmented into distinct frequency components: low-frequency components (smoothed image) are processed separately from high-frequency components (residual image). The sharpening operation is applied only to the low-frequency component, while the high-frequency component containing noise is preserved separately and combined later, thus avoiding noise amplification while maintaining edge sharpness.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

A smoothed version of the image serves as an intermediary representation that contains the structural information without the noise. This intermediary is used as the basis for sharpening operations, allowing the original image's noise-free structure to be enhanced without introducing additional noise amplification.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Object-affected harmful factors

If noise suppression via denoising is applied prior to sharpening, then noise amplification is reduced, but image details and natural appearance are lost

Engineering Contradiction:

Improvenoise amplificationVSAvoidimage details

Core Design Contradiction:

Object-affected harmful factorsVSLoss of information

Solution Approach 1:

Different processing qualities are applied to different frequency regions: aggressive smoothing is applied to low-frequency components where noise is less problematic, while high-frequency components are preserved with minimal processing to maintain natural image details and texture information. This selective processing prevents both noise amplification and detail loss.

Inventive Principle:

Principle #3Local quality

3Manufacturing precision

If image sharpening is applied to enhance local contrast, then edge definition is improved, but blur effects are introduced

Engineering Contradiction:

Improveedge definitionVSAvoidimage blur

Core Design Contradiction:

Manufacturing precisionVSShape

Solution Approach 1:

The processing approach dynamically adapts to different image regions by using adaptive kernels that adjust their behavior based on local image characteristics. In edge regions, the processing enhances definition, while in homogeneous regions, it maintains natural appearance, thus achieving edge enhancement without introducing blur effects.

Inventive Principle:

Principle #15Dynamics

Data Source

PatentUS8849057B2Methods for digital image sharpening with noise amplification avoidance

Publication Date: 2014.09.30 SIGMA CORP

AI summary

A method for sharpening an original digital image includes generating a smoothed image from the original image. A residual image is generated from the smoothed image and the original image. A sharpened smoothed image is generated. The residual image and the sharpened smoothed image are combined to produce a sharpened version of the original image.