Adaptive Correlation Block Size for Motion Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing motion estimation techniques for image sensors, such as those used in optical mice, face challenges in balancing precision, image capture frequency, and hardware costs, particularly in accurately determining mouse displacement speeds while minimizing the size of sensor networks and image capture frequency.
Innovation Solution
Adapting the size of the correlation block used for motion vector estimation based on historical motion vectors, allowing for dynamic adjustment of block dimensions in response to changes in mouse displacement speed, thereby optimizing motion vector determination with reduced hardware and processing requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a fixed-size correlation block is used for motion estimation, then the processing is simple and fast, but the measurement precision of motion vectors deteriorates when mouse displacement speed varies
Solution Approach 1:
The patent applies dynamics by making the correlation block size adaptive rather than fixed. The block dimensions are dynamically adjusted based on the magnitude of motion vectors from previous frames. When motion is intense, smaller blocks are used to maintain estimation accuracy, while during gentle motion, larger blocks provide better precision. This dynamic adaptation resolves the contradiction between maintaining high measurement precision across varying speeds and keeping processing complexity manageable.
Solution Approach 2:
The patent changes the parameter of correlation block size based on motion intensity. By monitoring the magnitude of motion vectors from previous frames and adjusting the block dimensions accordingly, the system optimizes measurement precision for different displacement speeds. This parameter adaptation allows the system to maintain high accuracy without requiring complex fixed-precision processing for all motion conditions.
2Measurement precision
If a large correlation block is used to improve motion estimation accuracy, then measurement precision improves, but the device complexity and processing requirements increase
Solution Approach 1:
The system dynamically adjusts correlation block size based on actual motion conditions rather than using a consistently large block. During periods of intense motion, smaller blocks are employed to maintain accuracy while reducing processing load. During gentle motion, larger blocks are used to maximize precision. This dynamic approach eliminates the need for permanently complex processing infrastructure required by fixed large blocks.
Solution Approach 2:
The patent changes the correlation block size parameter adaptively based on motion vector magnitude from previous frames. This parameter modulation allows the system to achieve high measurement precision when needed while reducing processing requirements during low-motion periods, resolving the contradiction between accuracy and complexity.
3Measurement precision
If image capture frequency is increased to accurately determine motion at high speeds, then measurement precision improves, but the use of energy and device complexity increase
Solution Approach 1:
The patent changes the correlation block size parameter based on detected motion intensity from previous frames. When high displacement speeds are detected, smaller blocks are used to maintain estimation accuracy without requiring increased capture frequency. During low-motion periods, larger blocks are employed to maximize precision at the original capture rate. This parameter adaptation allows accurate high-speed motion determination while maintaining constant, lower energy consumption.
Data Source
AI summary
A process for determining the displacement of an entity equipped with a sensor for capturing a sequence of images, comprising a step for determining a motion vector associated with a current image as a function of at least one correlation calculation between a first block of pixels in the current image and a second block of pixels from which the vector points towards said first block of pixels, with said second block being in a previous image in the sequence of images, wherein the dimensions of the first block are determined as a function of at least a motion vector associated with a previous image in the image sequence.