Spatially Varying Unsharp Mask Filter for Adaptive Noise Reduction

Solution Overview

Problem

Conventional unsharp masking techniques enhance noise in images by applying a uniform filter, leading to reduced image quality due to spatially varying noise across the image.

Innovation Solution

A spatially varying unsharp mask noise reduction filter that generates low-pass and high-pass filtered images and blends them using a shaping function, with the blending process controlled by a linear interpolation technique based on the high-pass signal's amplitude, to reduce noise while enhancing high-frequency information.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If a uniform filter kernel is applied to the entire image, then the filtering process is simple and fast, but noise is enhanced in regions where noise amplitude exceeds the threshold

Engineering Contradiction:

Improvefiltering speedVSAvoidimage quality

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The patent applies different filter kernel sizes based on the spatial location and noise characteristics of different regions in the image. The filter kernel size is dynamically adjusted according to the local noise amplitude and image content, allowing aggressive filtering in low-frequency regions while preserving detail in high-frequency regions, thus preventing noise enhancement while maintaining filtering effectiveness

Inventive Principle:

Principle #3Local quality

Solution Approach 2:

The patent introduces dynamic adaptation by adjusting the filter kernel size based on local image characteristics and noise levels. The filtering parameters are not fixed but are dynamically modified for each region, enabling the filter to respond to varying noise conditions across different parts of the image and avoid the noise enhancement problem associated with uniform filtering

Inventive Principle:

Principle #15Dynamics

2Reliability

If a large filter kernel is used to reduce noise, then noise reduction effectiveness increases, but edge information and high-frequency details are lost

Engineering Contradiction:

Improvenoise reduction effectivenessVSAvoidedge information

Core Design Contradiction:

ReliabilityVSLoss of information

Solution Approach 1:

The patent applies different filter kernel sizes based on the spatial location and noise characteristics of different regions in the image. The filter kernel size is dynamically adjusted according to the local noise amplitude and image content, allowing aggressive filtering in low-frequency regions while preserving detail in high-frequency regions, thus preventing noise enhancement while maintaining filtering effectiveness

Inventive Principle:

Principle #3Local quality

Solution Approach 2:

The patent effectively segments the image processing into different regions based on noise characteristics and frequency content. By identifying and treating low-frequency regions separately from high-frequency regions, the filter can apply stronger noise reduction where appropriate while preserving edge information in regions where it is critical

Inventive Principle:

Principle #1Segmentation

Data Source

PatentUS9082180B2System, method, and computer program product for implementing a spatially varying unsharp mask noise reduction filter

Publication Date: 2015.07.14 NVIDIA CORP

AI summary

A system, method, and computer program product for applying a spatially varying unsharp mask noise reduction filter is disclosed. The spatially varying unsharp mask noise reduction filter generates a low-pass filtered image by applying a low-pass filter to a digital image, generates a high-pass filtered image of the digital image, and generates an unsharp masked image based on the low-pass filtered image and the high-pass filtered image. The filter also blends the low-pass filtered image with the unsharp masked image based on a shaping function.