Resampling Position Calculation for Video Upsampling Quality

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing video compression techniques face significant degradation in performance at low bit rates, leading to substantial quality loss and artifacts in reconstructed video due to quantization and lossy processing, especially in high-frequency information, which is not adequately addressed by current spatially scalable codecs with inflexible and computationally expensive upsampling methods.

Innovation Solution

The development of high-accuracy position calculation techniques for picture resizing in spatially-scalable video coding and decoding, utilizing resampling scale factors and advanced filtering methods like the Mitchell-Netravali filter design, which provides flexible resampling ratios and alignments, improving upsampling quality while reducing computational complexity and memory requirements.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If conventional quantization and lossy processing are used in video compression, then bit rate is reduced, but video quality degrades significantly with artifacts appearing in reconstructed video

Engineering Contradiction:
Improvebit rateVSAvoidvideo quality
Core Design Contradiction:
Quantity of substanceVSManufacturing precision

Solution Approach 1:

The patent segments the video signal into different frequency components using transform coding (e.g., DCT), allowing selective processing of high-frequency information. By dividing the frequency spectrum into bands, the encoder can apply different quantization strategies to preserve important high-frequency details while maintaining compression efficiency.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent dynamically adjusts quantization parameters based on local image characteristics, motion activity, and frequency content. By changing quantization step sizes adaptively rather than uniformly, the system preserves high-frequency information where needed while achieving better overall compression, reducing artifacts in reconstructed video.

Inventive Principle:
Principle #35Parameter changes

2Quantity of substance

If spatially scalable video coding with downsampling is used, then bit rate is reduced, but high-frequency information is lost and quality degrades

Engineering Contradiction:
Improvebit rateVSAvoidhigh-frequency information
Core Design Contradiction:
Quantity of substanceVSLoss of information

Solution Approach 1:

The patent applies preliminary high-pass filtering and edge detection before downsampling to identify and preserve important high-frequency structures. By preparing the signal in advance with enhanced high-frequency components, the subsequent downsampling operation retains more useful information and produces better quality reconstructed video at lower bit rates.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent introduces an intermediary processing stage between encoding and decoding that includes adaptive filtering and detail enhancement. This intermediary layer reconstructs lost high-frequency information by analyzing patterns in the compressed data and applying appropriate synthesis filters, acting as a bridge that recovers information lost during compression.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Productivity

If traditional upsampling filters are used in decoder, then computational complexity is reduced, but upsampling quality is insufficient and artifacts remain

Engineering Contradiction:
Improvedecoding speedVSAvoidupsampling quality
Core Design Contradiction:
ProductivityVSManufacturing precision

Solution Approach 1:

The patent implements dynamic upsampling filter selection that adapts to the content being processed. Different filter kernels are chosen based on local image characteristics, motion vectors, and frequency content. This dynamic approach maintains high upsampling quality where needed while using simpler filters for uniform regions, optimizing the balance between quality and computational complexity.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent employs a nested multi-stage upsampling architecture where coarse upsampling is performed first using simple filters, followed by progressive refinement stages that add detail using more complex filtering. This nested approach achieves high final quality while keeping individual processing stages computationally manageable, maintaining decoding speed.

Inventive Principle:
Principle #7Nested doll (Nesting)

Data Source

PatentEP1985120B1Resampling and picture resizing operations for multi-resolution video coding and decoding
Publication Date: 2019.09.04 MICROSOFT TECHNOLOGY LICENSING LLC
  • EP1985120B1 patent drawingFigure 1
  • EP1985120B1 patent drawingFigure 2
  • EP1985120B1 patent drawingFigure 3

AI summary

Techniques and tools for high accuracy position calculation for picture resizing in applications such as spatially-scalable video coding and decoding are described. In one aspect, resampling of a video picture is performed according to a resampling scale factor. The resampling comprises computation of a sample value at a position i, j in a resampled array. The computation includes computing a derived horizontal or vertical sub-sample position x or y in a manner that involves approximating a value in part by multiplying a 2n value byan inverse (approximate or exact) of the upsampling scale factor. The approximating can be a rounding or some other kind of approximating, such as a ceiling or floor function that approximates to a nearby integer. The sample value is interpolated using a filter.