Spatio-Temporal Noise Reduction Controller Edge Slope Blending

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Aggressive compression modes in video processing introduce errors that reduce the effectiveness of temporal noise reduction, leading to biased noise reduction towards spatial methods and resolution loss in still images.

Innovation Solution

A spatio-temporal noise reduction controller that determines current and previous image edge slopes to adaptively blend images, selectively nullifying motion noise and adjusting gain values based on noise levels and motion status, thereby enhancing noise reduction accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If aggressive compression modes such as REMPEG are performed to reduce memory size, then memory size is reduced, but error is introduced that creates false motion detection and reduces temporal noise reduction effectiveness

Engineering Contradiction:

Improvememory sizeVSAvoidtemporal noise reduction effectiveness

Core Design Contradiction:

Quantity of substanceVSReliability

Solution Approach 1:

The patent applies preliminary action by performing spatial noise reduction on the previous frame before it is used for temporal noise reduction. This pre-processing step removes compression-induced noise and false motion artifacts from the previous frame, ensuring that when temporal noise reduction compares the current frame with the previous frame, the comparison is based on cleaner data. This resolves the contradiction by maintaining TNR effectiveness despite using compressed previous frames stored in limited memory.

Inventive Principle:

Principle #10Preliminary action

2Quantity of substance

If aggressive compression modes are used to store previous frames, then memory requirements are reduced, but resolution is lost and noise levels increase in still images

Engineering Contradiction:

Improvememory requirementsVSAvoidimage resolution

Core Design Contradiction:

Quantity of substanceVSManufacturing precision

Solution Approach 1:

By applying spatial noise reduction to the compressed previous frame before using it in temporal noise reduction, the patent restores image quality and reduces noise artifacts introduced by compression. This pre-processing step recovers resolution and reduces noise levels in still image regions, allowing the system to use aggressively compressed frames in memory while maintaining final output quality.

Inventive Principle:

Principle #10Preliminary action

3Reliability

If temporal noise reduction is reduced due to false motion detection, then false motion artifacts are reduced, but noise reduction becomes biased towards spatial methods only

Engineering Contradiction:

Improvemotion detection accuracyVSAvoidnoise reduction methodology

Core Design Contradiction:

ReliabilityVSAdaptability or versatility

Solution Approach 1:

The patent restores the balance between spatial and temporal noise reduction by pre-processing the previous frame with spatial noise reduction. This ensures that when temporal noise reduction operates, it works with cleaned data that doesn't contain false motion artifacts from compression. As a result, TNR can be effectively applied again without being biased solely toward spatial methods, maintaining methodological versatility.

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS8774549B2Compression error handling for temporal noise reduction

Publication Date: 2014.07.08 STMICROELECTRONICS INT NV

AI summary

A video processor includes a spatio-temporal noise reduction controller to determine current and previous image edge slopes and adaptively control a spatio-temporal noise reduction processor to blend current and previous images dependent on the current and previous image edge slope values.