Kernel Regression Image Reconstruction with Rotation and Reliability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing high-resolution image reconstruction methods, such as kernel regression and Motion-Assisted Steering Kernel (MASK) regression, struggle with complex motion and rotational object motion in multiple frames, leading to performance degradation and inability to adequately handle these motions.
Innovation Solution
The method incorporates a kernel function that accounts for registration reliability and rotation information, using adaptive weighting of registration residues and multi-scale prediction errors, and employs an affine transform to derive a rotation matrix for improved motion handling, thereby enhancing the accuracy of high-resolution image reconstruction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional kernel regression or MASK regression is used for high-resolution image reconstruction, then the method can handle simple motion, but it fails to adequately handle complex motion and rotational object motion, leading to performance degradation
Solution Approach 1:
The patent modifies the kernel function by incorporating rotation information and registration reliability as additional parameters. The kernel function is changed from a standard form to one that includes terms for rotation angle and registration quality, allowing the system to adapt to complex motions while maintaining reconstruction accuracy through parameter-driven adjustments.
Solution Approach 2:
The patent introduces dynamic adaptation of the kernel function based on local motion characteristics. By calculating registration reliability and rotation information for each local region, the system dynamically adjusts the kernel function to match the actual motion patterns, enabling it to handle both simple and complex motions effectively throughout different regions of the image.
2Measurement precision
If adaptive weighting of registration residues and multi-scale prediction errors is employed, then reconstruction accuracy improves, but computational complexity increases
Solution Approach 1:
The patent divides the image into multiple local regions and performs registration and kernel function calculation independently for each region. This segmentation allows the computationally intensive adaptive weighting to be applied locally rather than globally, reducing overall computational complexity while maintaining high reconstruction accuracy in each region.
Solution Approach 2:
The patent applies adaptive weighting and multi-scale prediction error calculation selectively in regions where complex motion is detected, rather than uniformly across the entire image. By identifying regions with high registration residue or complex motion patterns and applying enhanced processing only there, the system achieves high reconstruction accuracy where needed while minimizing unnecessary computational overhead in simpler regions.
Data Source
AI summary
A method and apparatus for reconstructing a high-resolution image based on multiple low-resolution images are disclosed. The method and apparatus incorporating an embodiment according to the present invention reconstructs the high-resolution image based on a kernel regression method using a modified kernel function. The kernel function takes into consideration of registration reliability of regression residue and rotational motion within the multiple low-resolution images. The registration reliability adjusts weighting on the regression residues according to local gradient estimated between neighboring values. Furthermore, multi-scale regression residue is used to alleviate impact of noise.