Adaptive Regularization Parameter for Medical Image Denoising
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Medical imaging techniques, such as PET and CT, often produce noisy images that can hinder diagnosis due to the influence of noise in the reconstructed images, making it challenging to accurately interpret and analyze the data.
Innovation Solution
A system and method that utilize a processor to determine a regularization parameter based on scanning data, including components like noise equivalent counts, spatial sensitivity of the scanner, and subject features, to generate a denoised image through an objective function, which incorporates total variation, quadratic, or Huber functions for improved image reconstruction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If medical imaging reconstruction is performed using conventional methods, then image data is obtained, but noise appears in the reconstructed image which influences image quality and causes difficulties in diagnosis
Solution Approach 1:
The patent applies parameter changes by dynamically adjusting the regularization parameter based on multiple components including noise equivalent counts characterizing data quality, spatial sensitivity values characterizing the scanner, and attenuation information characterizing subject features. This adaptive parameter adjustment optimizes the balance between noise suppression and image detail preservation, resolving the contradiction between reducing noise and maintaining measurement precision in medical images
Solution Approach 2:
The patent implements feedback mechanisms by using the determined regularization parameter in an iterative reconstruction process based on an objective function. The reconstruction process continuously refines the image by comparing reconstructed data with actual scanning data and adjusting parameters accordingly, providing feedback that progressively reduces noise while maintaining diagnostic quality
2Adaptability or versatility
If a fixed regularization parameter is used for image reconstruction, then the reconstruction process is simple, but it cannot adapt to different scanning data quality, scanner characteristics, or subject features
Solution Approach 1:
The patent segments the regularization parameter into multiple independent components: a first component characterizing noise equivalent counts (data quality), a second component characterizing spatial sensitivity values (scanner features), and a third component characterizing attenuation information (subject features). This segmentation allows each component to be determined and optimized independently, then combined to form the complete regularization parameter, achieving adaptability without excessive complexity
Solution Approach 2:
The patent creates a universal regularization parameter determination method that can be applied across different medical imaging scenarios by incorporating multiple characterizing components. The same framework adapts to various scanning conditions, scanner types, and subject features through its multi-component structure, providing universal applicability while maintaining the ability to handle specific case variations
Data Source
AI summary
The present disclosure provides a system and method for medical imaging. The method may include obtaining scanning data of a subject acquired using a scanner. The method may also include determining a reconstruction parameter of an objective function based at least in part on the scanning data, wherein the reconstruction parameter includes one or more global components that control an image quality over an entire image to be reconstructed and one or more local components that control the image quality over a portion of the image to be reconstructed. The method may further include generating the image of the subject by reconstructing the scanning data based on the objective function.