Adaptive Dequantization in Video Coding

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing video coding technologies face limitations in coding efficiency due to side information related to quantization parameters, which restricts the improvement of coding efficiency.

Innovation Solution

An adaptive dequantization method that estimates statistical features of the background area using spatial neighborhood information to derive a quantization adjustment factor, allowing flexible adjustment of dequantization processing without requiring additional bit overhead for transmitting quantization adjustment information.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If quantization parameter information is transmitted in the bitstream to adjust quantization steps, then coding efficiency is improved, but bit rate increases due to additional side information

Engineering Contradiction:
Improvecoding efficiencyVSAvoidbit rate
Core Design Contradiction:
ProductivityVSQuantity of substance

Solution Approach 1:

The decoder performs dequantization adjustment by itself using locally derived statistics (variance and mean of neighboring blocks) without requiring any quantization adjustment information to be transmitted from the encoder. The system serves itself by using available decoded information to adapt the dequantization process, eliminating the need for additional bitstream data.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent extracts only the essential information (variance and mean of neighboring blocks) that is already available at the decoder side, and uses this extracted information to derive the quantization adjustment factor. This avoids transmitting redundant quantization parameter information while still achieving adaptive dequantization.

Inventive Principle:
Principle #2Taking out (Extraction)

2Device complexity

If fixed dequantization is used to simplify processing, then device complexity is reduced, but coding efficiency deteriorates due to inability to adapt to different visual importance areas

Engineering Contradiction:
Improvedequantization processing complexityVSAvoidcoding efficiency
Core Design Contradiction:
Device complexityVSProductivity

Solution Approach 1:

The patent applies different dequantization adjustments to different transform blocks based on their local characteristics. By calculating the variance and mean of neighboring blocks for each transform block, the system adapts the quantization adjustment factor locally, allowing important areas (low variance) to retain more detail while less important areas (high variance) use coarser quantization.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The patent dynamically changes the dequantization parameter (quantization adjustment factor) based on the statistical characteristics of the image content. The adjustment factor is derived from the variance and mean of neighboring blocks, allowing the system to adaptively modify dequantization behavior without complex processing structures.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP3163875B1Adaptive inverse-quantization method and apparatus in video coding
Publication Date: 2020.11.25 HUAWEI TECH CO LTD
  • EP3163875B1 patent drawingFigure 1~2
  • EP3163875B1 patent drawingFigure 3~4
  • EP3163875B1 patent drawingFigure 5

AI summary

The present invention discloses an adaptive dequantization method and apparatus in video coding. According to the method provided in the present invention, a statistical feature of a background area in which a current transform block is located is estimated by using spatial neighborhood information, or temporal neighborhood information, or spatial neighborhood information and temporal neighborhood information of the transform block, a quantization adjustment factor for dequantization is adaptively derived, and dequantization processing is flexibly adjusted. In comparison with an existing solution in which quantization adjustment information is transmitted in a bitstream, according to the method and the apparatus that are provided in the present invention, extra bit overheads are not required to transmit the quantization adjustment information, thereby further improving coding efficiency.