Digital Picture Noise Reduction Combining High and Low Noise Processed Pictures

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing noise reduction techniques for digital pictures often result in excessive blurriness or unnatural appearance due to aggressive noise reduction, which can eliminate high-frequency content and make smooth areas appear flat.

Innovation Solution

A method involving two parallel noise reduction processes: one applying a strong noise reduction technique and the other a weak or no noise reduction, with the combined output using filtering and weighted averaging to balance noise reduction and maintain natural picture quality.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If aggressive noise reduction is applied to reduce low-frequency noise, then noise reduction level is improved, but high-frequency content is lost making the picture look unnatural

Engineering Contradiction:

Improvenoise reduction levelVSAvoidhigh-frequency content

Core Design Contradiction:

Manufacturing precisionVSLoss of information

Solution Approach 1:

The patent segments the noise reduction process into two distinct processing paths: one path applies aggressive noise reduction to reduce low-frequency noise, while the other path preserves high-frequency content with minimal noise reduction. These segmented paths are then recombined to produce the final output, allowing simultaneous achievement of noise reduction and detail preservation.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent applies different noise reduction strengths to different frequency components and picture regions. Aggressive noise reduction is applied primarily to low-frequency components and smooth areas, while high-frequency components and edge regions receive minimal or no noise reduction, preserving local quality and natural appearance.

Inventive Principle:

Principle #3Local quality

2Manufacturing precision

If aggressive noise reduction is applied, then smooth areas appear less noisy, but smooth areas become completely flat and unnatural

Engineering Contradiction:

Improvenoise reduction levelVSAvoidsmooth area texture

Core Design Contradiction:

Manufacturing precisionVSShape

Solution Approach 1:

The processing is segmented into multiple pathways where one pathway applies strong noise reduction to smooth areas while another pathway preserves the original texture information. These pathways are then combined using blending techniques that maintain smooth area noise reduction while preserving subtle texture characteristics.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent applies noise reduction selectively and partially rather than uniformly across the entire image. By applying noise reduction only where needed (in smooth areas) and at controlled levels, it achieves noise reduction without completely flattening the texture, maintaining natural appearance.

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS9041836B1Digital picture noise reduction by combining high-noise and low-noise processed pictures

Publication Date: 2015.05.26 AMBARELLA INT LP

AI summary

A method for processing a digital picture is disclosed. The method may include steps (A) to (C). Step (A) may generate a first picture by processing the digital picture using a first noise reduction technique in a circuit. Step (B) may generate a second picture by processing the digital picture using a second noise reduction technique. The first noise reduction technique may achieve a higher noise reduction than the second noise reduction technique. Step (C) may generate an output picture by combining the first picture and the second picture.