True Motion Estimation Using Extended-Block FFT and Phase Correlation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video motion estimation algorithms focus on video compression rather than true motion between frames, leading to inefficiencies in picture quality improvement, particularly due to high computational complexity and limited ability to account for object motion in and out of blocks.
Innovation Solution
The method employs local motion estimation using extended-block FFTs, phase correlation to determine motion vector candidates, and a hierarchical variable block search with predictive motion vector tracking to identify true motion, minimizing computational costs and accounting for object motion across different block sizes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If optical flow approach is used to estimate true motion, then measurement precision of motion is improved, but device complexity and computational power requirements increase significantly
Solution Approach 1:
The patent segments the image into blocks and processes each block independently using phase correlation, rather than processing every pixel as in optical flow. This segmentation reduces computational complexity while maintaining motion estimation accuracy for each block.
Solution Approach 2:
The patent replaces the complex mechanical/optical flow computation with phase correlation in the frequency domain using FFT. This substitution transforms the computational approach from spatial domain pixel-by-pixel analysis to frequency domain block-based analysis, significantly reducing complexity.
2Device complexity
If block matching algorithm is used for motion estimation, then device complexity is reduced for real-time processing, but measurement precision of true motion deteriorates
Solution Approach 1:
The patent substitutes block matching with phase correlation based on FFT. This replacement maintains the block-based approach for simplicity but uses frequency domain correlation which provides superior motion estimation accuracy by capturing phase information and handling various motion types more effectively.
Solution Approach 2:
The patent changes the computational domain from spatial domain (block matching) to frequency domain (phase correlation with FFT). This parameter change enables more accurate motion estimation while maintaining computational efficiency through the properties of FFT and phase correlation.
3Productivity
If existing motion search algorithms are applied directly for video compression, then productivity is improved, but measurement precision of true motion deteriorates
Solution Approach 1:
The patent replaces conventional motion search algorithms with phase correlation based on FFT. This substitution provides more accurate true motion estimation which improves video quality enhancement while maintaining productivity through efficient frequency domain computation.
Solution Approach 2:
The patent changes the motion estimation approach from compression-oriented block matching to true motion-oriented phase correlation. This parameter change enables accurate motion detection for quality improvement applications while maintaining computational efficiency suitable for practical use.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach effectively estimates true motion with reduced computational complexity, enabling improved picture quality by focusing on object motion rather than pixel motion, and providing translation, scaling, and rotation capabilities beyond conventional motion estimators.
Implementation Method 1
An extended-block FFT is calculated for each block
Implementation Method 2
The aim of the phase correlation process is to measure the movement of objects between two scenes by measuring the correlation of the pixels in the current block against the pixels in the delayed reference block
Data Source
AI summary
Local motion estimation is described herein. Each picture of a video is partitioned into blocks for the local motion estimation. An extended-block FFT is calculated for each block, where the extended-block denotes that a certain area around the block is also included for applying FFT. Extending the block for FFT helps to account for the motion of objects that are moving into or out of the block. Phase correlation is applied to attain a set of Motion Vector (MV) candidates for the blocks, and a cost function is evaluated for each MV. If no MV candidate produces a cost function below a pre-defined threshold, a hierarchical variable block matching search is applied and the process is repeated with blocks for finer resolution. Also, predictive MV candidates are used during the block matching search along with temporal constraints tracking to select an MV that yields the minimum cost function.