Adaptive Motion Vector Generation for Video Processing

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing motion estimation methods in video processing, such as full search, 3-step search, and 3-D recursive search, face challenges with slow convergence rates, uncontrollable motion estimation, and inefficiency in true motion estimation due to high computational intensity and random factors.

Innovation Solution

An integrated circuit with an image information generator, an adaptive motion vector candidate generator, and a block matching unit that determines image information and adjusts search range sizes or generates updated motion vector candidates based on this information to improve motion estimation efficiency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If full search method is used for motion estimation, then measurement precision of motion vector is improved, but computational complexity and processing time increase significantly

Engineering Contradiction:

Improvemotion vector estimation accuracyVSAvoidcomputational complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The search range is divided into multiple sub-regions, and motion estimation is performed separately in each sub-region. This segmentation allows the algorithm to maintain high precision by searching local areas thoroughly while reducing overall computational complexity by avoiding exhaustive search across the entire search window.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The algorithm performs preliminary motion estimation using a reduced search set to identify candidate motion vectors before conducting detailed block matching. This preliminary action narrows down the search space, allowing subsequent precision estimation to focus only on promising candidates rather than all possible vectors.

Inventive Principle:

Principle #10Preliminary action

2Speed

If 3-D recursive search with random factor is used, then convergence speed is improved, but motion estimation becomes uncontrollable

Engineering Contradiction:

Improveconvergence speedVSAvoidmotion estimation controllability

Core Design Contradiction:

SpeedVSReliability

Solution Approach 1:

The algorithm dynamically adjusts the search strategy based on image content characteristics. For regions with simple motion, a faster recursive search is used, while regions with complex or random motion patterns trigger a more thorough block matching approach, making the process adaptive rather than purely random or fixed.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The algorithm uses feedback from preliminary motion estimation results to control subsequent detailed matching. If the preliminary estimation shows good correlation, the algorithm proceeds with confidence; if not, it triggers expanded searching, making the process controllable rather than purely random.

Inventive Principle:

Principle #23Feedback

3Device complexity

If fixed scenario search method is used, then device complexity is reduced, but adaptability to true motion estimation deteriorates

Engineering Contradiction:

Improvealgorithm simplicityVSAvoidmotion estimation adaptability

Core Design Contradiction:

Device complexityVSAdaptability or versatility

Solution Approach 1:

The algorithm transitions from fixed scenario search to dynamic adaptive search by incorporating image content analysis. The search strategy is adjusted based on detected motion characteristics, edge information, and correlation metrics, allowing the simple framework to adapt to various motion patterns without requiring complex predetermined scenarios.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The algorithm changes search parameters such as search range, block size, and matching threshold based on image content analysis. This allows a relatively simple algorithm structure to achieve high adaptability by dynamically adjusting its operating parameters rather than being constrained to fixed scenarios.

Inventive Principle:

Principle #35Parameter changes

4Measurement precision

If exhaustive block matching is performed for all MV candidates, then measurement precision is improved, but processing time increases

Engineering Contradiction:

Improvemotion vector accuracyVSAvoidprocessing time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The algorithm performs exhaustive block matching only for a selected subset of promising motion vector candidates rather than all possible candidates. By using preliminary estimation to identify the most likely candidates, the algorithm achieves high precision for the critical cases while saving time by avoiding exhaustive search for unlikely candidates.

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS9641861B2Method and integrated circuit for video processing

Publication Date: 2017.05.02 XUESHAN TECH INC

AI summary

An integrated circuit capable of video processing and a method thereof is disclosed. The integrated circuit comprises an image information generator, an adaptive motion vector (MV) candidate generator, and a block matching unit. The image information generator receives reference and current video data to determine image information. The adaptive MV candidate generator, coupled to the image information generator, receives the reference and current video data to generate spatial and temporal MV candidates, and generates an updated MV candidate based on the image information. The block matching unit, coupled to the adaptive MV candidate generator, receives the spatial and temporal MV candidates and the updated MV candidate and performs block matching to determine an output MV therefrom.