Deblocking Filter Strength Adjustment for Extended Macroblocks

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Blocking artifacts increase when using extended macroblock sizes, particularly at low bit rates, leading to insufficient removal of deblocking artifacts and poorer image quality with conventional filters.

Innovation Solution

An image processing apparatus and method that adjusts filter strength based on neighboring block sizes at block boundaries, applying filtering to decoded image data and adjusting parameter values for improved deblocking filter performance, especially for larger block sizes.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If extended macroblock sizes are used to increase compression rate, then encoding efficiency is improved, but blocking artifacts increase and image quality deteriorates

Engineering Contradiction:
Improvecompression rateVSAvoidblocking artifacts
Core Design Contradiction:
ProductivityVSObject-affected harmful factors

Solution Approach 1:

The patent applies different filter strengths to different block boundaries based on the sizes of adjacent blocks. Specifically, stronger filtering is applied to boundaries between large blocks (e.g., 32×32 or 64×64 macroblocks) where blocking artifacts are more prominent, while weaker filtering is applied to boundaries between smaller blocks. This local adaptation of filter strength effectively reduces blocking artifacts in regions where they occur most severely while maintaining overall image quality.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The patent dynamically adjusts the deblocking filter strength parameter based on the macroblock sizes at block boundaries. The filter strength is modified according to the specific combination of adjacent block sizes, with larger blocks receiving stronger filtering. This parameter adaptation allows the system to maintain high compression rates with extended macroblock sizes while compensating for the increased blocking artifacts through targeted filter strength adjustment.

Inventive Principle:
Principle #35Parameter changes

2Device complexity

If conventional deblocking filter is used with extended macroblock sizes, then device complexity is maintained, but filtering performance is insufficient and image quality deteriorates

Engineering Contradiction:
Improvefilter structureVSAvoiddeblocking filter performance
Core Design Contradiction:
Device complexityVSManufacturing precision

Solution Approach 1:

The patent introduces dynamic adjustment of filter strength based on block size characteristics. Rather than using a fixed filter strength for all block boundaries, the system dynamically selects and applies different filter strengths according to the sizes of adjacent blocks. This dynamic adaptation maintains the simplicity of the overall filter structure while significantly improving deblocking performance for extended macroblock sizes.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentUS10373295B2Image processing apparatus and image processing method
Publication Date: 2019.08.06 SONY GROUP CORP
  • US10373295B2 patent drawing
  • US10373295B2 patent drawing
  • US10373295B2 patent drawing

AI summary

It is configured such that decoded images with favorable image quality are obtained, even in the case of using extended macroblock sizes.A deblocking filter 24 applies filtering to reduce blocking artifacts from an encoded stream in which image data has been encoded in individual blocks. A filter strength adjuster 41 adjusts the filter strength according to block sizes of the predictive image data. For example, during image encoding, the filter strength is adjusted according to the block size of the predictive image data yielding the best encoding efficiency generated using filtered and decoded image data. It may also be configured such that filtering is applied to remove blocking artifacts from decoded image data during image decoding, with the filter strength being adjusted according to the block sizes of the predictive image data.