Slice-Level Intra Block Copy for Video Coding Efficiency

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing video coding techniques lack finer control over intra block copy (IBC) modes, which limits the accuracy of prediction and coding efficiency, especially in scenarios with text and artificial graphics where repeated patterns exist.

Innovation Solution

The proposed techniques enable IBC on a slice level, allowing for a two-dimensional vector to identify predictive blocks within a broader intended region of the picture, including above, above-right, above-left, and left areas, rather than being limited to immediate neighbors, thereby enhancing prediction accuracy and coding efficiency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If IBC is enabled at the picture level, then coding efficiency is improved for scenarios with repeated patterns, but control over IBC usage is too coarse and cannot be optimized for specific regions

Engineering Contradiction:
Improvecoding efficiencyVSAvoidcontrol granularity
Core Design Contradiction:
ProductivityVSAdaptability or versatility

Solution Approach 1:

The picture is divided into multiple slices, and IBC control is applied at the slice level rather than the entire picture level. Each slice can independently enable or disable IBC based on its content characteristics, allowing fine-grained control while maintaining coding efficiency where beneficial.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Different slices can have different IBC enablement states based on their local content characteristics. Slices containing text or artificial graphics with repeated patterns can enable IBC, while other slices can disable it, optimizing performance locally rather than uniformly across the entire picture.

Inventive Principle:
Principle #3Local quality

2Device complexity

If reference blocks are limited to immediate neighboring blocks, then decoding complexity is reduced, but prediction accuracy deteriorates for blocks farther apart with similar patterns

Engineering Contradiction:
Improvedecoding complexityVSAvoidprediction accuracy
Core Design Contradiction:
Device complexityVSMeasurement precision

Solution Approach 1:

The reference block search is extended from only horizontal/vertical immediate neighbors to include diagonal directions (above-right, above-left). This dimensional expansion allows the system to access predictive blocks that are spatially farther away but still maintain pattern similarity, improving prediction accuracy without significantly increasing decoding complexity.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Solution Approach 2:

The reference picture buffer serves multiple purposes: it stores not only immediate neighboring blocks for simple prediction but also blocks from extended regions (diagonal directions) for more accurate prediction when patterns repeat across larger distances. This multi-functional use of the buffer improves prediction versatility.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Data Source

PatentEP3304908B1Slice level intra block copy
Publication Date: 2020.10.28 QUALCOMM INC
  • EP3304908B1 patent drawingFigure 1
  • EP3304908B1 patent drawingFigure 2
  • EP3304908B1 patent drawingFigure 3

AI summary

A device for decoding video data includes a memory configured to store the video data and one or more processors configured to receive a slice of the video data, parse an intra block copy (IBC) syntax element to determine that an IBC mode is enabled for the slice, parse a slice type syntax element associated with the slice to determine the slice is an I slice, and decode the slice as an I slice by decoding all blocks of the slice using intra prediction coding modes.