Adaptive Thresholding for Signal Reconstruction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional compressed sensing techniques for image/signal reconstruction are computationally intensive, require substantial processing time, and involve manual intervention due to their iterative and non-linear nature, which is inefficient and laborious, especially in applications like MRI and CT imaging.
Innovation Solution
A method that initializes two image solutions and a linear combination solution, iteratively updates and adaptively thresholds transform coefficients to generate selected coefficients for reconstructing data signals with reduced error and noise, thereby reducing processing time and manual intervention.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If conventional compressed sensing techniques are used for image reconstruction, then image quality can be achieved with fewer data samples, but processing time and computational complexity increase substantially
Solution Approach 1:
The patent segments the image reconstruction process into multiple stages: initial image formation, transform coefficient calculation, adaptive thresholding, and iterative refinement. This segmentation allows each stage to process specific portions of the data independently, reducing overall computational burden while maintaining reconstruction quality.
Solution Approach 2:
The patent applies preliminary actions by pre-calculating transform coefficients and establishing initial image solutions before the main iterative reconstruction process. This preliminary processing reduces the computational load during subsequent iterations and accelerates convergence to the final reconstructed image.
2Measurement precision
If iterative methods with non-linear cost functions are used for compressed sensing, then reconstruction accuracy improves, but computational intensity and processing time increase
Solution Approach 1:
The patent implements feedback mechanisms where the algorithm continuously evaluates reconstruction quality and adjusts parameters accordingly. The adaptive thresholding process uses feedback from error metrics to dynamically modify reconstruction parameters, improving accuracy while controlling computational complexity through automated parameter adjustment.
Solution Approach 2:
The patent changes parameters dynamically during the reconstruction process, particularly through adaptive thresholding where threshold values are adjusted based on error metrics and iteration progress. This parameter adaptation allows the algorithm to optimize reconstruction accuracy at each stage while managing computational resources more efficiently.
3Quantity of substance
If conventional compressed sensing methods are applied, then data compression is achieved, but manual intervention and customization are required
Solution Approach 1:
The patent implements self-service through automated parameter adjustment and adaptive thresholding that occurs without manual intervention. The algorithm automatically optimizes reconstruction parameters based on the input data characteristics, eliminating the need for user customization while maintaining optimal compression and reconstruction performance.
4Quantity of substance
If fewer data samples are used for reconstruction, then storage space and bandwidth requirements are reduced, but reconstruction quality may deteriorate
Solution Approach 1:
The patent replaces traditional mechanical sampling approaches with compressed sensing techniques that use mathematical transforms and iterative optimization. This substitution allows high-quality reconstruction from fewer data samples by leveraging the compressible nature of medical images through transform-domain processing and adaptive thresholding.
Data Source
AI summary
A signal processing method is presented. The method includes acquiring undersampled data corresponding to an object, initializing a first image solution and a second image solution, determining a linear combination solution based upon the first image solution and the second image solution, generating a plurality of selected coefficients by iteratively updating the first image solution, the second image solution and the linear combination solution and adaptively thresholding one or more transform coefficients utilizing the undersampled data, an updated first image solution, an updated second image solution and an updated linear combination solution, and reconstructing a data signal using the plurality of selected coefficients.