Low-Dose PET Restoration With Joint PET-MR Dictionary Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing low-dose PET images suffer from significant noise and loss of details, leading to poor restoration quality, complex restoration processes, and low accuracy, which affects the efficiency and effectiveness of PET/MR imaging.
Innovation Solution
A low-dose PET image restoration method involving blocking processing, sparse coding, and dictionary updating to construct a joint dictionary, utilizing K-means clustering and K-Singular Value Decomposition (K-SVD) for iterative updates, and mapping to restore images to standard-dose quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If low-dose PET imaging is used to reduce radiation exposure, then radiation dose is reduced, but image quality deteriorates due to increased noise and loss of details
Solution Approach 1:
The patent introduces a joint dictionary as an intermediary structure that combines information from PET and MR images. This joint dictionary serves as a mediator to transfer structural information from the high-quality MR image to the low-dose PET image during restoration, enabling quality improvement without increasing radiation dose.
Solution Approach 2:
The patent merges PET and MR image information into a unified joint dictionary through sparse coding and iterative updates. By combining the complementary strengths of both modalities (PET's metabolic information and MR's structural detail), the system achieves superior restoration quality compared to single-modality approaches.
2Reliability
If traditional image restoration methods are applied to low-dose PET images, then some noise reduction is achieved, but restoration accuracy and quality remain poor
Solution Approach 1:
The patent changes the fundamental parameters of image restoration by introducing a joint dictionary with adaptive basis vectors that are iteratively updated. This transforms the restoration process from fixed-parameter methods to adaptive, data-driven parameter adjustment, significantly improving restoration accuracy.
Solution Approach 2:
The patent implements an iterative feedback mechanism where the joint dictionary is continuously updated based on the residual error between the restored PET image and the original low-dose PET image. This feedback loop enables progressive refinement of restoration accuracy until convergence criteria are met.
3Device complexity
If conventional restoration algorithms are used, then restoration process is simple, but restoration time is long and efficiency is low
Solution Approach 1:
The patent performs preliminary actions by pre-computing the joint dictionary from training data before actual image restoration. This pre-processing step establishes a reusable dictionary that can be applied to multiple images, significantly reducing the computational burden and time required for restoration of individual images.
Solution Approach 2:
The patent segments the restoration process into distinct phases: joint dictionary construction from training data, sparse coding of the low-dose PET image, iterative dictionary updates, and final image reconstruction. This segmentation allows each phase to be optimized independently and enables parallel processing of multiple images.
Data Source
AI summary
A low-dose PET image restoration method and system, a device, and a medium are provided. The method includes: S1, performing blocking processing on a training image comprising a low-dose PET image, an MR image, and a standard-dose PET image to obtain a first patch, and performing first preprocessing on the first patch to obtain a second patch; S2, obtaining, according to the second patch, a first joint dictionary by means of sparse coding and dictionary updating; and S3, restoring the low-dose PET image to a restored image of the standard-dose PET image according to the first joint dictionary.