Adaptive Image Noise Filtering for Blur and Ghosting Control

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing noise reduction technologies for images fail to consider the intensity of side effects caused by noise filters, leading to issues like blurring, resolution degradation, ghosting, and motion blur, particularly in stationary and moving objects.

Innovation Solution

An image noise reduction device and method using machine learning to measure and adjust noise filter intensity based on the side effects caused by noise filtering, employing a 3D noise filter for stationary objects and a 2D noise filter for moving objects, with AI parameters determined through supervised learning on image attributes and side effect differences.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If the intensity of the noise filter is increased to improve noise removal performance, then the noise reduction effect is improved, but side effects such as blurring, resolution degradation, ghosting, and motion blur increase

Engineering Contradiction:

Improvenoise removal performanceVSAvoidside effects (blurring, ghosting, motion blur)

Core Design Contradiction:

Manufacturing precisionVSObject-affected harmful factors

Solution Approach 1:

The patent applies dynamics by making the noise filter intensity adaptive rather than fixed. The system dynamically adjusts the intensity of the noise filter based on real-time analysis of image characteristics, object motion status, and side effect measurements. This allows the filter intensity to vary across different regions and time points, optimizing noise removal while minimizing blurring, ghosting, and motion blur artifacts.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The patent implements parameter changes by modifying the noise filter intensity parameter based on measured side effects. The system continuously monitors side effect intensity and adjusts the noise filter parameters accordingly - reducing intensity when side effects are detected and increasing it when noise removal is prioritized. This dynamic parameter adjustment resolves the contradiction between noise removal effectiveness and side effect minimization.

Inventive Principle:

Principle #35Parameter changes

2Manufacturing precision

If 3D noise filtering is applied to stationary objects to improve noise removal, then noise reduction performance is improved, but ghosting and dragging effects occur on moving objects

Engineering Contradiction:

Improvenoise reduction performanceVSAvoidghosting and dragging effects

Core Design Contradiction:

Manufacturing precisionVSObject-affected harmful factors

Solution Approach 1:

The patent applies local quality by differentiating noise filtering treatment between stationary and moving objects within the same image. The system identifies moving objects through motion detection and applies appropriate filtering strategies locally - using 2D spatial filtering for moving objects to avoid ghosting, and 3D temporal filtering for stationary backgrounds where ghosting is not an issue. This localized approach allows optimal noise removal for each object type without cross-contamination of artifacts.

Inventive Principle:

Principle #3Local quality

Solution Approach 2:

The patent implements segmentation by separating the image processing into distinct regions based on motion status. The system divides the image into moving object regions and stationary background regions, then applies different noise filtering algorithms to each segment. This segmentation enables the system to apply 3D filtering only where appropriate (stationary regions) and 2D filtering where motion artifacts would occur (moving object regions), resolving the ghosting problem.

Inventive Principle:

Principle #1Segmentation

3Object-affected harmful factors

If 2D noise filtering is applied to moving objects to reduce ghosting, then ghosting effects are minimized, but noise removal performance deteriorates compared to 3D filtering

Engineering Contradiction:

Improveghosting effectsVSAvoidnoise removal performance

Core Design Contradiction:

Object-affected harmful factorsVSManufacturing precision

Solution Approach 1:

The patent applies merging by combining multiple noise filtering approaches (2D spatial filtering and 3D temporal filtering) into a unified adaptive system. Rather than choosing one filter type exclusively, the system merges both filtering methodologies and dynamically selects or combines them based on local image characteristics, object motion status, and measured side effects. This combination allows the system to achieve both ghosting reduction and effective noise removal by utilizing the strengths of each filtering approach where appropriate.

Inventive Principle:

Principle #5Merging (Combining)

Data Source

PatentUS12482218B2Machine learning-based image noise learning server and image noise reduction device

Publication Date: 2025.11.25 HANWHA VISION CO LTD

AI summary

An image noise learning server includes an image input interface configured to receive training images, and at least one processor configured to control an image extractor to extract, from the training images, a first image including a stationary object and a second image including a moving object, a noise filter to obtain a third image by applying noise filtering with a first intensity to the second image, the third image including the moving object, a labeling unit to determine an intensity of a side effect based on a difference between the stationary object included in the first image and the moving object included in the third image, and a machine learning unit to receive, as a label, the determined intensity of the side effect and image attributes of the training images, and obtain artificial intelligence (AI) parameters by performing machine learning on the second image based on the received label.