Adaptive Digital Image Noise Reduction via Per-Pixel Filtering

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing video compression techniques, such as block-based compression standards like AVC/H.264, introduce artifacts like block noise and mosquito noise into digital images, which can degrade image quality, especially in low bit rate data and large display formats.

Innovation Solution

A method and apparatus that determine mosquito and block noise on a per-pixel basis, using adaptive filters to reduce compression artifacts by combining filter outputs based on noise values, employing edge adaptive filtering to preserve true image features and selectively apply filtering to areas with noise, thereby improving image quality.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If block-based compression standards like AVC/H.264 are used to compress video content, then storage area needed and transmission time are reduced, but block noise and mosquito noise artifacts are introduced into the image data

Engineering Contradiction:
Improvecompression efficiencyVSAvoidcompression artifacts
Core Design Contradiction:
ProductivityVSObject-generated harmful factors

Solution Approach 1:

The image is divided into blocks corresponding to the compression standard's block structure. Noise detection and filtering operations are performed independently on each block, allowing selective processing of noisy regions while preserving quality in clean regions.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Different filtering strengths and types are applied to different regions of the image based on local noise characteristics. The filtering operation adapts its parameters locally to reduce artifacts in noisy areas while maintaining image quality in areas without artifacts.

Inventive Principle:
Principle #3Local quality

2Object-generated harmful factors

If strong filtering is applied to reduce compression artifacts, then noise is reduced, but image details and true features are blurred or lost

Engineering Contradiction:
Improvenoise reductionVSAvoidimage detail preservation
Core Design Contradiction:
Object-generated harmful factorsVSManufacturing precision

Solution Approach 1:

The filtering operation is made dynamic and adaptive rather than static. Filtering parameters such as strength, type, and application region are adjusted dynamically based on detected noise characteristics in different areas of the image, allowing the filter to be strong where needed and gentle where not needed.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The filtering strength and type vary locally across the image based on detected artifact presence. Regions with compression artifacts receive stronger filtering while regions without artifacts receive minimal or no filtering, preserving true image features in clean regions.

Inventive Principle:
Principle #3Local quality

3Object-generated harmful factors

If traditional noise reduction methods are applied uniformly across the entire image, then overall noise is reduced, but non-noisy areas are unnecessarily smoothed

Engineering Contradiction:
Improveoverall noise reductionVSAvoidselective area preservation
Core Design Contradiction:
Object-generated harmful factorsVSManufacturing precision

Solution Approach 1:

The image is segmented into regions with artifacts and regions without artifacts through block-level noise detection. This segmentation allows the filtering operation to be selectively applied only to noisy regions, leaving clean regions untouched.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The filtering operation adapts its behavior locally across different image regions. Some regions receive strong filtering while others receive no filtering, based on the presence or absence of compression artifacts in each region.

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP2327219B1Reducing digital image noise
Publication Date: 2016.11.09 MARVELL WORLD TRADE LTD
  • EP2327219B1 patent drawingFigure 1
  • EP2327219B1 patent drawingFigure 2
  • EP2327219B1 patent drawingFigure 3

AI summary

Devices, systems, methods, and other embodiments associated with reducing digital image noise are described, in one embodiment, a method includes determining, on a per pixel basis, mosquito noise values associated with pixels of a digital image. The method determines, on a per pixel basis, block noise values associated with the digital image. The method filters the digital image with a plurality of adaptive filters. A compression artifact in the digital image is reduced. The compression artifact is reduced by combining filter outputs from the plurality of adaptive filters. The filter outputs are combined based, at least in part, on the mosquito noise values and the block noise values.