Wide-Angle Video Encoding via Reference Block Rotation and Scaling

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional methods for encoding moving image data of wide viewing angle images face challenges with increased residual and coding bit rate due to deformation, particularly when using motion prediction by translating block-shaped data.

Innovation Solution

A coding apparatus that divides wide viewing angle image data into blocks, extracts motion-compensated reference blocks, performs rotation and/or scaling adjustments on these blocks, calculates residual signals, and encodes them, while inserting adjustment processing information with motion vectors into the encoded stream.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If motion prediction is performed by translating block-shaped data according to a conventional method, then the encoding process is simple, but the residual becomes large and the coding bit rate increases

Engineering Contradiction:

Improveencoding process complexityVSAvoidresidual

Core Design Contradiction:

Device complexityVSLoss of information

Solution Approach 1:

The patent applies rotation and scaling transformations to the reference block to match the geometric deformation characteristics of wide viewing angle images. By changing the parameters (rotation angle, scaling factor) of the reference block, the prediction accuracy is improved, reducing the residual error between the current block and predicted block, thereby lowering the coding bit rate while maintaining encoding complexity at an acceptable level.

Inventive Principle:

Principle #35Parameter changes

2Device complexity

If motion prediction is performed by translating block-shaped data according to a conventional method, then the encoding process is simple, but the coding bit rate increases

Engineering Contradiction:

Improveencoding process complexityVSAvoidcoding bit rate

Core Design Contradiction:

Device complexityVSQuantity of substance

Solution Approach 1:

Inventive Principle:

Principle #35Parameter changes

3Loss of information

If rotation and scaling adjustment processing is performed on the reference block, then the residual and coding bit rate are reduced, but the device complexity increases

Engineering Contradiction:

ImproveresidualVSAvoidencoding process complexity

Core Design Contradiction:

Loss of informationVSDevice complexity

Solution Approach 1:

The patent performs rotation and scaling adjustments on the reference block before extracting motion-compensated reference blocks. This preliminary action prepares the reference block in advance to match the expected geometric transformation, making the subsequent motion compensation more accurate and reducing the residual error, while the increased complexity is concentrated in the preprocessing stage.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

Inventive Principle:

Principle #35Parameter changes

4Quantity of substance

If rotation and scaling adjustment processing is performed on the reference block, then the residual and coding bit rate are reduced, but the device complexity increases

Engineering Contradiction:

Improvecoding bit rateVSAvoidencoding process complexity

Core Design Contradiction:

Quantity of substanceVSDevice complexity

Solution Approach 1:

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentEP3518534B1Encoding device, encoding method, decoding device, decoding method, transmission device, and reception device

Publication Date: 2023.05.17 SONY GROUP CORP

Figure 1~2
Figure 3
Figure 4

AI summary

Moving image data of a wide viewing angle image is satisfactorily encoded. Moving image data of a wide viewing angle image is divided into blocks to obtain a coding target block. A motion-compensated reference block is extracted for each coding target block. The reference block is subjected to adjustment processing of rotation and/or scaling. A residual signal is calculated on the basis of a pixel value of the coding target block and a pixel value of the reference block subjected to the adjustment processing for each coding target block, and the residual signal of the coding target block is coded to obtain an encoded stream. Adjustment processing information is inserted into the encoded stream together with the motion vector for each coding target block. The residual can be reduced, the coding bit rate can be reduced, and consequently the coding performance of motion prediction can be improved.