Motion Vector Candidate List Construction for Video Encoding Efficiency

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional motion compensation prediction technologies face inefficiencies in encoding motion information, particularly when the motion of a processing target block differs from that of neighboring blocks or blocks in the same position of another encoded picture, leading to increased code amounts and reduced encoding efficiency.

Innovation Solution

A picture encoding device and method that constructs a candidate list of motion information from neighboring blocks, deriving and combining motion information from multiple prediction lists to generate new candidates, which are then encoded, thereby optimizing the encoding of motion vectors and reducing the need for transmitting coding vectors.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If the block size of motion compensation prediction is made finer and more diverse, then the precision of motion prediction is improved, but the number of motion vectors increases leading to increased code amount

Engineering Contradiction:

Improvemotion prediction precisionVSAvoidcode amount of coding vectors

Core Design Contradiction:

Measurement precisionVSQuantity of substance

Solution Approach 1:

The patent combines motion information from multiple neighboring blocks (left, upper, upper-right blocks) to construct a candidate list for motion vector prediction. By merging motion vectors from these different sources and selecting the best candidate, the system achieves accurate motion prediction without needing to transmit all individual motion vectors, thus reducing code amount while maintaining prediction precision.

Inventive Principle:

Principle #5Merging (Combining)

Solution Approach 2:

The patent performs preliminary construction of a motion information candidate list before actual motion compensation prediction. Motion vectors from neighboring blocks are pre-collected and organized into candidates, allowing the system to efficiently select the most appropriate motion vector without exhaustive search, thereby reducing the code amount required for transmitting motion information.

Inventive Principle:

Principle #10Preliminary action

2Productivity

If motion information of neighboring blocks is used without transmitting coding vector, then encoding efficiency is improved, but accuracy decreases when motion of processing target block differs from neighboring blocks

Engineering Contradiction:

Improveencoding efficiencyVSAvoidmotion prediction accuracy

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The patent dynamically constructs a motion information candidate list based on the specific characteristics of neighboring blocks. Instead of using a fixed motion vector from a single neighboring block, the system adaptively selects from multiple candidates (including left, upper, and upper-right blocks) the one that best matches the processing target block's motion, thereby maintaining high accuracy while improving encoding efficiency.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The patent changes the parameters of motion information by collecting motion vectors from multiple neighboring blocks with different spatial positions and characteristics. By varying the source of motion information across different candidates in the candidate list, the system can adapt to different motion scenarios and maintain prediction accuracy even when the processing target block's motion differs from any single neighboring block.

Inventive Principle:

Principle #35Parameter changes

3Quantity of substance

If a single vector predictor is used, then code amount is reduced, but precision of motion prediction deteriorates

Engineering Contradiction:

Improvecode amount of coding vectorsVSAvoidvector predictor precision

Core Design Contradiction:

Quantity of substanceVSMeasurement precision

Solution Approach 1:

The patent merges motion information from multiple neighboring blocks (left block, upper block, upper-right block) to construct a candidate list containing multiple motion vector candidates. By combining information from these different sources rather than relying on a single vector predictor, the system achieves more precise motion prediction while still transmitting only one selected candidate, thus maintaining low code amount.

Inventive Principle:

Principle #5Merging (Combining)

Solution Approach 2:

The patent introduces a motion information candidate list as an intermediary structure between the raw motion vectors of neighboring blocks and the final selected motion vector. This candidate list serves as a mediator that organizes and evaluates multiple motion information sources, allowing the system to select the best predictor without transmitting all intermediate data, thereby balancing precision and code amount.

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS10009624B2Picture encoding device, picture encoding method, picture encoding program, picture decoding device, picture decoding method, and picture decoding program

Publication Date: 2018.06.26 JVC KENWOOD CORP

AI summary

A candidate list construction unit selects a plurality of blocks each having one or two pieces of motion information containing at least information about a motion vector and information about a reference picture from a plurality of neighboring encoded blocks of an encoding target block and constructs a candidate list containing candidates of the motion information used for the motion compensation prediction from the motion information of the selected blocks. A selected candidate generator generates a new candidate of the motion information by combining the motion information of a first prediction list derived by the first motion information deriving unit and the motion information of the second prediction list derived by a second motion information deriving unit.