Adaptive Joint Sparse Coding for Parallel MRI Reconstruction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing parallel magnetic resonance imaging methods face challenges in maintaining reconstruction quality as acceleration multiple increases, particularly with calibration-free reconstruction methods that rely solely on joint sparsity in transformation domains, leading to sensitivity to initial values and complex calculations.
Innovation Solution
The method employs adaptive joint sparse codes based on block sparsity, constructing a calibration-free parallel magnetic resonance imaging model using a reconstructed model defined by minimizing a specific objective function, updating the dictionary and sparse coefficients through gradient descent, and iteratively refining K-space data and images to enhance sparsity and reconstruction quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If calibration-free reconstruction method is used, then robustness against sensitivity information prediction errors is improved, but reconstruction quality deteriorates when acceleration multiple increases
Solution Approach 1:
The patent changes the mathematical parameters of the reconstruction model by introducing a mixed norm (combining l2 and l2,1 norms) instead of using only traditional l2 norm. This parameter change in the objective function enables the model to maintain robustness while achieving better reconstruction quality at high acceleration multiples by enforcing joint sparsity across multiple channels
Solution Approach 2:
The patent creates a composite reconstruction approach by combining multiple mathematical concepts: calibration-free reconstruction, joint sparsity constraint, and mixed norm regularization. This composite method integrates the advantages of different approaches to simultaneously achieve robustness and high reconstruction quality
2Productivity
If acceleration multiple is increased, then imaging speed is improved, but reconstruction quality deteriorates
Solution Approach 1:
The patent moves from considering sparsity in a single transformation domain to enforcing joint sparsity across multiple channels (adding a dimensional aspect). By applying l2,1 norm that operates across channels, the method exploits the additional dimensional structure to maintain reconstruction quality at higher acceleration multiples
Solution Approach 2:
The patent modifies the regularization term in the objective function from traditional l2 norm to a mixed norm involving l2 and l2,1 norms. This parameter change enables the model to handle higher acceleration factors by enforcing a stronger sparsity constraint that leverages inter-channel correlations
3Device complexity
If joint sparsity is used only in transformation domain, then computational complexity is reduced, but sparsity is insufficient leading to degraded reconstruction quality
Solution Approach 1:
The patent merges the sparsity constraints from the transformation domain with joint sparsity constraints across multiple channels. By combining l2 norm (transformation domain sparsity) and l2,1 norm (joint channel sparsity) in a mixed norm objective function, the method achieves sufficient sparsity for high-quality reconstruction without excessive computational complexity
Data Source
AI summary
Provided are a parallel magnetic resonance imaging method and apparatus based on adaptive joint sparse codes and a computer-readable medium. The method includes solving an l2−lF−l2,1 minimization objective, where the l2 norm is a data fitting term, the lF norm is a sparse representation error, and the l2,1 mixed norm is the joint sparsity constraining across multiple channels; separately updating each of a sparse matrix, a dictionary and K-space data with a corresponding algorithm, and obtaining a reconstructed image by a sum of root mean squares of all the channels. The joint sparsity of the channels is developed using the norm l2,1. In this manner, calibration is not required while information sparsity is developed. Moreover, the method is robust.