Quincunx-Sampled Video Deblocking via Component Segmentation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for processing sampled and multiplexed image and video data face challenges in efficiently deblocking and interpolating quincunx-sampled content, leading to cross-view contamination and suboptimal compression efficiency due to the complexity of handling different sampling patterns and interleaving arrangements.

Innovation Solution

The proposed method involves de-multiplexing composite sampled data into component pictures, processing each component separately with deblocking, denoising, and filtering, while considering the spatial position of samples in the full resolution grid to avoid contamination and improve interpolation accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If quincunx-sampled content is processed using existing deblocking and interpolation methods, then processing can be performed on sampled data, but cross-view contamination occurs and compression efficiency is suboptimal

Engineering Contradiction:
Improvedeblocking performanceVSAvoidcross-view contamination
Core Design Contradiction:
ReliabilityVSObject-generated harmful factors

Solution Approach 1:

The patent segments the processing by separating even and odd sample lines into distinct component pictures, then processes each component independently through deblocking and interpolation operations. This segmentation prevents cross-contamination between different view components while maintaining processing effectiveness on quincunx-sampled content.

Inventive Principle:
Principle #1Segmentation

2Manufacturing precision

If different sampling patterns and interleaving arrangements are handled using complex processing methods, then accurate processing can be achieved, but computational complexity increases

Engineering Contradiction:
Improveinterpolation accuracyVSAvoidprocessing complexity
Core Design Contradiction:
Manufacturing precisionVSDevice complexity

Solution Approach 1:

The patent applies local quality by treating even and odd sample lines with different processing approaches appropriate to their specific characteristics. Each component picture undergoes processing tailored to its sampling pattern, enabling accurate interpolation while simplifying the overall computational burden compared to uniform complex processing.

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP2596635B1Enhancement methods for sampled and multiplexed image and video data
Publication Date: 2020.05.20 DOLBY LABORATORIES LICENSING CORP
  • EP2596635B1 patent drawingFigure 1
  • EP2596635B1 patent drawingFigure 2
  • EP2596635B1 patent drawingFigure 3

AI summary

Enhancement methods for sampled and multiplexed image and video data are described. Each component picture is separately processed either after de-multiplexing or on the fly. Processing and de-multiplexing can be combined in a single joint step. The methods apply to both encoding and decoding system and include applications to scalable video coding systems.