Beamforming Module Dimensionality Reduction for Ultrasonic Imaging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current ultrasonic imaging apparatuses face challenges in reducing the calculation load and time required for beamforming, as well as the resources used, which affects the efficiency and speed of generating high-quality ultrasonic images.
Innovation Solution
A beamforming module that includes a conversion unit, a weight calculator, and a synthesizer, which converts input signals using a conversion function to reduce dimensions and calculate optimal weights for beamforming, thereby reducing calculation load and time by employing a combination of basis vectors from principle component analysis and minimum variance techniques.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If adaptive beamforming is used to improve image quality, then measurement precision is improved, but calculation load increases
Solution Approach 1:
The beamforming process is segmented into two distinct stages: a preprocessing stage that computes transformation matrices (eigenvectors, principal components, or whitening matrices) from training data, and a real-time beamforming stage that applies these pre-computed transformations to incoming signals. This segmentation allows complex adaptive beamforming to be decomposed into offline preparation and online execution, significantly reducing real-time calculation load while preserving image quality.
Solution Approach 2:
The patent performs preliminary computations of transformation matrices using training data before actual beamforming operations. By pre-computing eigenvectors, principal components, or whitening matrices during a training phase, the system eliminates the need to perform these computationally intensive calculations during real-time imaging, thereby reducing calculation load during actual operation while maintaining adaptive beamforming performance.
2Productivity
If conventional beamforming methods are used, then device complexity is reduced, but productivity decreases
Solution Approach 1:
The beamforming process is segmented into a training phase for pre-computing transformation matrices and an execution phase for applying these matrices. This segmentation enables the system to achieve fast real-time processing (high productivity) by moving computationally intensive operations to the training phase, while keeping the real-time beamforming operation simple and efficient.
Solution Approach 2:
The patent creates a simplified model of the signal environment through transformation matrices derived from training data. Instead of performing complex adaptive beamforming calculations during real-time operation, the system uses these pre-computed transformation matrices as a copy or approximation of the optimal beamforming solution, enabling fast processing without sacrificing significant performance.
3Device complexity
If dimensionality reduction is applied, then calculation load is reduced, but measurement precision may deteriorate
Solution Approach 1:
The patent transforms the signal representation from the original high-dimensional channel space to a reduced-dimensional transformed space using pre-computed transformation matrices. This parameter change in the signal representation space, combined with the use of training data to optimize the transformation, maintains measurement precision while enabling dimensionality reduction that significantly lowers calculation load during real-time beamforming.
Data Source
AI summary
A beamforming module includes a conversion unit configured to convert an input signal to generate a converted signal using at least one conversion function, a weight calculator configured to calculate a converted signal weight as a weight for the converted signal, and a synthesizer configured to generate a result signal using the converted signal and the converted signal weight.