Probabilistic PET Image Reconstruction via Noise Model Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current medical image reconstruction techniques, particularly in positron emission tomography (PET), face challenges in enhancing image quality without introducing artifacts, which hinders accurate monitoring of tumor growth and diagnostic precision.
Innovation Solution
A method and device for medical image reconstruction that involves obtaining image data, a noise model, and an initial model indicative of expected properties, along with a mapping of the medical scanner, to determine a set of candidate images using a probabilistic approach, thereby improving image data quality and enabling more quantitative analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If image processing methods are used to increase image quality, then image quality is improved, but artifacts are introduced into the representation
Solution Approach 1:
The patent introduces a probabilistic model as an intermediary between the raw image data and the final reconstructed image. This model represents uncertainty and multiple possible states, allowing the system to explore various candidate images and select the most probable one, thereby improving quality without introducing deterministic artifacts
Solution Approach 2:
The patent changes the parameter representation from deterministic pixel values to probability distributions. By representing each pixel as a distribution of possible values rather than a single fixed value, the system can capture uncertainty and avoid committing to potentially erroneous reconstructions that would create artifacts
2Measurement precision
If conventional reconstruction techniques are used, then processing speed is maintained, but image quality and quantitative analysis capability are limited
Solution Approach 1:
The patent segments the reconstruction process into distinct stages: obtaining image data, obtaining a noise model, obtaining an initial model, determining candidate images, and determining the final representation. This segmentation allows each stage to be optimized independently and facilitates the integration of complex probabilistic modeling without overwhelming the overall system
Solution Approach 2:
The patent replaces traditional deterministic reconstruction algorithms with a probabilistic approach. Instead of using fixed mathematical transformations, the system employs probability distributions and statistical models to represent and process image data, enabling more accurate quantitative analysis through Bayesian inference and other probabilistic methods
Data Source
AI summary
Disclosed is a device and a method for medical image reconstruction. The method comprises obtaining, image data of a medical scanner; obtaining a noise model for the image data from the medical scanner; obtaining an initial model indicative of expected image data properties; obtaining a mapping, wherein the mapping is indicative of a mapping from the medical scanner; determining a set of candidate images based on the image data, the noise model, the initial model, and the mapping; and determining and outputting a first representation of the image data based on the set of candidate images.