Adaptive Video Post-Processing for Compression Artifact Removal

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Video codecs introduce noticeable compression artifacts during encoding, and conferencing tools often suffer from quality degradation due to network congestion or bandwidth limitations, leading to suboptimal video quality and increased bandwidth consumption.

Innovation Solution

Adaptive post-processing of decoded video using machine learning models that leverage scenario detection, segmentation, and video quality analysis to selectively apply post-processing operations, enhancing video quality without significantly increasing bandwidth or reducing quality.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If lossy compression is applied to reduce bandwidth consumption, then network bandwidth usage decreases, but video quality deteriorates with noticeable compression artifacts

Engineering Contradiction:

Improvebandwidth consumptionVSAvoidvideo quality

Core Design Contradiction:

Quantity of substanceVSManufacturing precision

Solution Approach 1:

The patent applies machine learning post-processing models to convert the harmful compression artifacts into beneficial visual outcomes. The models learn to reconstruct high-quality video from low-quality compressed inputs, effectively transforming the degradation caused by lossy compression into an opportunity for intelligent enhancement and artifact removal

Inventive Principle:

Principle #22Blessing in disguise (Convert harm into benefit)

Solution Approach 2:

The system changes the quality parameter of video encoding by using aggressive lossy compression to reduce bandwidth, then compensates by applying ML-based post-processing that reconstructs high-quality visual output from the compressed input, effectively decoupling the compression ratio from the final perceived quality

Inventive Principle:

Principle #35Parameter changes

2Reliability

If video is encoded at low quality to accommodate bandwidth limitations, then more packets are delivered timely, but extensive compression artifacts appear

Engineering Contradiction:

Improvepacket delivery reliabilityVSAvoidvideo quality

Core Design Contradiction:

ReliabilityVSManufacturing precision

Solution Approach 1:

The ML post-processing models transform the harmful compression artifacts resulting from low-quality encoding into beneficial visual quality. By training models on paired low-quality and high-quality video, the system learns to reconstruct details and remove artifacts, converting the degraded input into high-quality output

Inventive Principle:

Principle #22Blessing in disguise (Convert harm into benefit)

Solution Approach 2:

The machine learning model acts as an intermediary between the compressed video input and the final displayed output. This intermediary layer processes the degraded video signal, applying learned transformations to reconstruct high-quality visual content before presentation to the user

Inventive Principle:

Principle #24Intermediary (Mediator)

3Manufacturing precision

If multiple post-processing models are maintained for different scenarios, then video quality for specific scenarios improves, but device complexity increases

Engineering Contradiction:

Improvescenario-specific video qualityVSAvoidmodel selection complexity

Core Design Contradiction:

Manufacturing precisionVSDevice complexity

Solution Approach 1:

The patent segments the video processing task by dividing it into different scenario categories (e.g., talking head, conference room, presentation). Each segment is handled by a specialized post-processing model trained for that specific scenario, allowing the system to apply the most appropriate processing to each type of content while maintaining manageable complexity through organized categorization

Inventive Principle:

Principle #1Segmentation

Data Source

PatentUS20260006259A1Machine learning models for adaptive post-processing using results of scenario detection in conferencing tools

Publication Date: 2026.01.01 MICROSOFT TECHNOLOGY LICENSING LLC

AI summary

Innovations in machine learning (“ML”) models used in adaptive post-processing of decoded video in a conferencing tool are described. For example, as part of post-processing of decoded video, a super-resolution/video restoration model increases spatial resolution (e.g., by interpolation between sample values), mitigates compression artifacts, and mitigates upscaling artifacts introduced when increasing spatial resolution. Or, as another example, as part of post-processing of decoded video, a video restoration model mitigates compression artifacts, without increasing spatial resolution. For adaptive post-processing, a post-processing model can be selectively applied depending on results of scenario detection, results of segmentation, and/or results of video quality analysis. With the innovations, a conferencing tool can in effect provide video at higher quality without significantly increasing the network bandwidth consumed by the video or, alternatively, provide video using less network bandwidth without significantly hurting the quality of the video.