Motion Estimator Using Candidate Vector Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing motion estimation methods often incorrectly estimate motion vectors, leading to increased motion blur and inaccurate interpolation frames during frame rate conversion, as they may select vectors that do not represent real motion.
Innovation Solution
A motion estimator that calculates multiple candidate motion vectors for each block based on previous motion vectors, using averaging and block matching algorithms to determine a final motion vector with minimal prediction error and vector difference, thereby generating a more accurate interpolation frame.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a single motion vector is selected based on minimal motion prediction error, then the estimation process is simple and fast, but the selected vector may not represent real motion leading to inaccuracies
Solution Approach 1:
The patent segments the motion estimation process into multiple stages: first calculating candidate motion vectors based on previous motion vectors, then refining them by calculating motion prediction errors and selecting the optimal one. This segmentation allows the system to maintain speed while improving accuracy through systematic evaluation of multiple candidates rather than relying on a single potentially incorrect vector.
Solution Approach 2:
The patent performs preliminary calculations of multiple candidate motion vectors before final selection. By pre-calculating several candidate vectors and their corresponding motion prediction errors, the system prepares multiple options in advance, then selects the best one based on the minimal error criterion. This preliminary action ensures that the final selection is based on comprehensive evaluation rather than single-point estimation.
2Productivity
If candidate motion vectors are calculated based on previous motion vectors, then the estimation process is efficient, but wrong vectors may be propagated leading to increased motion blur
Solution Approach 1:
The patent implements feedback by calculating motion prediction errors for each candidate motion vector and using these errors to evaluate and select the optimal vector. The motion prediction error serves as feedback information that indicates how well a candidate vector represents actual motion, allowing the system to reject wrong vectors and propagate only accurate motion information, thereby reducing motion blur while maintaining efficiency.
Solution Approach 2:
The patent changes the parameter used for evaluation from simple vector magnitude to motion prediction error. By using motion prediction error as the selection criterion instead of relying solely on previous motion vector values, the system can identify and select vectors that truly represent motion, preventing the propagation of wrong vectors that cause motion blur while maintaining computational efficiency.
3Measurement precision
If multiple candidate motion vectors are evaluated, then the accuracy of motion vector selection is improved, but the computational complexity increases
Solution Approach 1:
The patent segments the motion estimation process into distinct functional units: a first candidate motion vector calculating unit that generates initial candidates, and a second candidate motion vector calculating unit that refines them by evaluating motion prediction errors. This segmentation organizes the complexity into manageable stages, improving accuracy through systematic evaluation while keeping the process structure clear and manageable.
Solution Approach 2:
The patent performs preliminary calculation of motion prediction errors for all candidate vectors before final selection. By pre-computing the error values for multiple candidates, the system enables comprehensive evaluation of each vector's accuracy, ensuring the selection of the most accurate motion vector while organizing the computational complexity into a systematic pre-evaluation process.
Data Source
AI summary
A motion estimator for generating an interpolation frame based on current and previous frames is provided. The motion estimator includes: a first candidate motion vector calculating unit which divides the current frame into plural blocks and generates plural first candidate motion vector groups, each group including a first candidate motion vector for each of the blocks calculated based on a previous motion vector previously estimated; a second candidate motion vector calculating unit which calculates plural second candidate motion vectors for each of the blocks, each of the second candidate motion vectors being calculated based on a respective first candidate motion vector included in each of the first candidate motion vector group; and a final motion vector deciding unit which selects a final motion vector for each of the blocks from among the second candidate motion vectors, the interpolation frame being generated based on the final motion vector.