Video Encoding Reference Vector Target Frame Selection

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional video encoding methods face inefficiencies when encoding motion vectors or disparity vectors, especially when the reference frame used for prediction differs between adjacent blocks, leading to increased code amounts due to larger differences between predicted and actual vectors.

Innovation Solution

A video encoding method that divides the image into areas, generates a predicted image using image information from multiple frames, and encodes differential information between the encoding target area and the predicted image, using a unified reference vector target frame to reduce code amount by unifying primary image variation factors.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If a predicted vector is generated using motion vectors from adjacent blocks with different reference frames, then the prediction process can be simplified, but the difference between predicted and actual motion vectors increases leading to larger code amounts

Engineering Contradiction:

Improveprediction process complexityVSAvoidcode amount

Core Design Contradiction:

Device complexityVSLoss of information

Solution Approach 1:

The patent segments the prediction process into two distinct components: reference frame selection and motion vector prediction. By separating these functions, the system can select the most appropriate reference frame based on the block's characteristics before generating the predicted vector, thereby maintaining prediction accuracy while managing complexity.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent performs preliminary reference frame selection before motion vector prediction. By determining the optimal reference frame in advance based on block characteristics and temporal correlations, the system ensures that subsequent motion vector predictions are made from the most suitable reference, reducing the difference between predicted and actual vectors.

Inventive Principle:

Principle #10Preliminary action

2Measurement precision

If different reference frames are used for adjacent blocks to optimize local prediction accuracy, then prediction quality improves, but the amount of code required to encode motion vectors increases

Engineering Contradiction:

Improveprediction accuracyVSAvoidcode amount

Core Design Contradiction:

Measurement precisionVSLoss of information

Solution Approach 1:

The patent applies local quality by allowing different reference frames to be selected for different blocks based on their specific characteristics. Each block can independently choose the reference frame that best matches its motion characteristics, achieving optimal local prediction accuracy while the overall system manages code efficiency through standardized encoding procedures.

Inventive Principle:

Principle #3Local quality

Solution Approach 2:

The patent changes the reference frame parameter dynamically for each block based on motion characteristics and temporal correlations. By adapting the reference frame selection to local motion patterns, the system achieves higher prediction accuracy without proportionally increasing code requirements, as the reference frame indices can be efficiently encoded.

Inventive Principle:

Principle #35Parameter changes

3Adaptability or versatility

If motion vectors are encoded using difference from predicted vectors without unified reference frame, then encoding flexibility is maintained, but encoding efficiency decreases due to larger vector differences

Engineering Contradiction:

Improveencoding flexibilityVSAvoidencoding efficiency

Core Design Contradiction:

Adaptability or versatilityVSProductivity

Solution Approach 1:

The patent introduces dynamics by making the reference frame selection adaptive rather than static. The system dynamically selects reference frames based on block characteristics, motion patterns, and temporal correlations, allowing the encoding process to adapt to varying content requirements while maintaining efficient code representation through standardized differential encoding.

Inventive Principle:

Principle #15Dynamics

Data Source

PatentUS8654854B2Video encoding method and decoding method, apparatuses therefor, programs therefor, and storage media which store the programs

Publication Date: 2014.02.18 NIPPON TELEGRAPH & TELEPHONE CORP

AI summary

A video encoding method includes selecting a reference vector target frame and a reference frame from among already-encoded frames; encoding information for designating each frame; setting a reference vector for indicating an area in the reference vector target frame with respect to an encoding target area; encoding the reference vector; performing a corresponding area search by using image information of a reference vector target area, which belongs to the reference vector target frame and is indicated by the reference vector, and the reference frame; determining a reference area in the reference frame based on the search result; generating a predicted image by using image information of the reference frame, which corresponds to the reference area; and encoding differential information between image information of the encoding target area and the predicted image.