Baseline Estimation With Green's Function Convolution for Deblurred Signals
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods fail to effectively remove out-of-focus contributions and noise components in input signal data, such as images, which are characterized by blurred artifacts and noise similar to out-of-focus effects, caused by factors like aliasing, noisy backgrounds, non-linear absorption, and moving targets.
Innovation Solution
A computer-implemented method using least-square minimization with a penalty term to separate in-focus and out-of-focus components in input signal data, employing polynomial or spline fits, and a half-quadratic minimization scheme to compute baseline estimation, followed by subtraction to obtain deblurred output signal data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If standard image sharpening methods such as deconvolution are applied, then in-focus features are enhanced, but out-of-focus artifacts remain and cannot be removed
Solution Approach 1:
The image is segmented into in-focus and out-of-focus components through mathematical decomposition. The sharp image I(x) is expressed as the sum of a blurred component B(x) and a sharp component S(x), allowing separate processing of each component to remove out-of-focus artifacts while preserving in-focus features
Solution Approach 2:
A point spread function h(x) serves as an intermediary element that models the blurring process. By introducing this mathematical mediator representing the optical system's characteristics, the patent enables the separation and removal of out-of-focus contributions through deconvolution operations
2Object-affected harmful factors
If polynomial fit is applied to spectral data for background subtraction, then baseline is removed, but noise components with similar characteristics are not effectively separated
Solution Approach 1:
The patent changes the approach from simple polynomial fitting to a more sophisticated parameter-based separation method using point spread functions and sharp image constraints. By introducing parameters characterizing the optical system and the sharp component, the method achieves more reliable separation of baseline and noise components
Data Source
Figure 1
Figure 2~3
Figure 4~6
AI summary
The invention relates to an apparatus and a method for obtaining baseline estimation data (f(x;)) of input signal data (l(x;)). Obtaining accurate baseline estimation data allows to efficiently remove noise fromt the input signal data, such as image, sonar, sound, ultrasound tomographic and/or seismographic data. The baseline estimation data are computed using a convolution with a Green's function. The convolution is computationally more efficient than the known matrix multiplication.