Motion Vector Merge Prediction with Compressed Motion Memory

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing image encoding methods, such as those in H.264, face challenges in reducing the code amount of motion information and reference frame numbers due to differences in predicted motion information derivation methods, leading to increased memory capacity and encoding inefficiency.

Innovation Solution

An image encoding apparatus that divides input images into pixel blocks, selects predicted motion information from a motion information buffer, and obtains representative motion information to reduce the amount of motion information stored, using a motion information compressor to minimize memory usage and enhance encoding efficiency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Volume of stationary object

If representative motion information is derived in a predetermined block and only the representative motion information is stored in the motion information memory, then the capacity of the motion information memory is reduced, but the temporal correlation of the predicted motion information reduces and the code amount of the motion information increases

Engineering Contradiction:

Improvemotion information memory capacityVSAvoidtemporal correlation of predicted motion information

Core Design Contradiction:

Volume of stationary objectVSLoss of information

Solution Approach 1:

The patent divides the prediction unit into multiple sub-blocks and derives motion information for each sub-block separately. This segmentation allows the system to maintain temporal correlation within each sub-block while reducing the overall memory capacity required, as only essential motion information for each segment is stored rather than complete motion information for the entire block.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent applies different motion information derivation methods to different regions or sub-blocks within the prediction unit. By optimizing motion information storage and prediction locally for each sub-block based on its specific characteristics, the system maintains high temporal correlation where needed while reducing memory usage in other areas.

Inventive Principle:

Principle #3Local quality

2Measurement precision

If motion information of a reference frame is used to derive predicted motion information of an encoding target block, then prediction accuracy is improved, but the capacity of the motion information memory increases

Engineering Contradiction:

Improveprediction accuracyVSAvoidmotion information memory capacity

Core Design Contradiction:

Measurement precisionVSVolume of stationary object

Solution Approach 1:

The patent extracts and stores only the essential motion information from reference frames that is most useful for predicting motion in the encoding target block. By selectively extracting only the necessary motion parameters and storing them in the motion information memory, the system maintains high prediction accuracy while minimizing memory capacity requirements.

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent performs preliminary processing of motion information from reference frames before storage, pre-selecting and organizing only the most relevant motion data. This preliminary action ensures that when prediction is needed, the system can quickly access accurate motion information without having to store and process all reference frame data, thus maintaining prediction accuracy with reduced memory usage.

Inventive Principle:

Principle #10Preliminary action

3Measurement precision

If multiple reference frames are used for motion compensation, then prediction accuracy is improved, but the code amount of motion information and reference frame numbers increases

Engineering Contradiction:

Improvemotion compensation accuracyVSAvoidcode amount of motion information

Core Design Contradiction:

Measurement precisionVSQuantity of substance

Solution Approach 1:

The patent uses a selective approach to reference frame utilization, employing multiple reference frames only when necessary for accurate prediction while using fewer or no reference frames when motion is minimal or predictable. This partial application of multi-reference-frame techniques reduces the code amount required for motion information while maintaining prediction accuracy where it matters most.

Inventive Principle:

Principle #16Partial or excessive action

Solution Approach 2:

The patent dynamically adjusts the number and type of reference frames used for motion compensation based on the specific characteristics of each prediction unit and its motion patterns. This dynamic adaptation allows the system to achieve high prediction accuracy when multiple reference frames are beneficial while minimizing code amount when fewer reference frames suffice.

Inventive Principle:

Principle #15Dynamics

Data Source

PatentUS12593062B2Image encoding and decoding method with merge flag and motion vectors

Publication Date: 2026.03.31 KK TOSHIBA

AI summary

A method of dividing an input image signal into pixel blocks, and performing inter-prediction on the divided pixel blocks. This method includes selecting predicted motion information from a motion information buffer storing motion information in an encoded region, and predicting motion information of an encoding target block by using the predicted motion information. The method further includes acquiring representative motion information from a plurality of items of motion information in an encoded region in accordance with first information indicating a method of selecting the predicted motion information, thereby obtaining only the representative motion information.