Automated Regularization Tuning in MRI Compressed Sensing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for selecting the regularization parameter β in compressed sensing (CS) MRI image reconstruction are computationally intensive and require iterative processes, leading to inefficiencies and potential errors in optimizing image quality due to the trade-off between noise and resolution.
Innovation Solution
An automated method is introduced to estimate the optimal regularization parameter β by applying a multi-scale transformation and thresholding technique, allowing for a single CS reconstruction with improved computational efficiency and reduced human interaction, thereby optimizing image quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If iterative reconstruction methods are used to recover images with lower sampling rate, then compressed sensing can achieve good image quality, but the process becomes slow and computationally expensive
Solution Approach 1:
The patent applies preliminary action by performing a multi-scale transformation on the initial image estimate before the iterative reconstruction process. This preliminary processing step prepares the image data in advance by transforming it into multiple frequency domains, which then guides the iterative reconstruction to converge faster and more efficiently, reducing the computational cost while maintaining image quality.
Solution Approach 2:
The patent segments the image reconstruction problem into multiple scales using wavelet or curvelet transforms. By dividing the image into different frequency components (approximation and detail coefficients at multiple scales), the method can process each scale separately and adaptively, which improves the efficiency of the iterative reconstruction and reduces overall computational complexity.
2Object-affected harmful factors
If more regularization is applied in CS reconstruction, then noise is reduced, but smoothing increases leading to blurring and reduced resolution
Solution Approach 1:
The patent applies local quality by using multi-scale transformations that allow different levels of regularization to be applied at different frequency scales. High-frequency components (details) can have stronger regularization to reduce noise, while low-frequency components (approximations) maintain weaker regularization to preserve resolution. This scale-dependent approach enables simultaneous noise reduction and resolution preservation.
Solution Approach 2:
The patent implements dynamics by making the regularization parameter adaptive rather than fixed. The regularization strength is dynamically adjusted based on the frequency content and scale of the image data at each iteration and each scale level. This allows the system to automatically optimize the balance between noise reduction and resolution preservation in real-time during the reconstruction process.
3Productivity
If automated methods are used to select regularization parameter, then computational efficiency improves, but the complexity of the method increases
Solution Approach 1:
The patent applies self-service by implementing an automated regularization parameter selection method that uses the image data itself to determine the optimal regularization strength. The multi-scale transformation and thresholding techniques allow the system to self-evaluate the image characteristics and automatically select appropriate parameters without external intervention, reducing the need for complex user input while maintaining computational efficiency.
Solution Approach 2:
The patent changes parameters by transforming the image into multiple frequency domains using wavelet or curvelet transforms, which creates a new parameter space where the regularization can be optimized. By working in the transformed domain rather than the original spatial domain, the method can more efficiently search for optimal regularization parameters and apply them selectively across different scales, improving computational efficiency.
Data Source
AI summary
An apparatus and method are provided to simultaneously provide good image quality and fast image reconstruction from magnetic resonance imaging (MRI) data by selecting an appropriate value for the regularization parameter used in compressed sensing (CS) image reconstruction. In CS reconstruction a high-resolution image can be reconstructed from randomized undersampled data by imposing sparsity in multi-scale transformation (e.g., wavelet) domain. Further, in the transformation domain, a threshold can be determined between signal and noise levels of the transform coefficients. A regularization parameter based on this threshold scales the regularization term, which imposes sparsity, relative to the data fidelity term in an objective function, thereby balancing the tradeoff between noise and smoothing.