Synthesizing High Dose CT Images via Probabilistic Voxel Modeling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
There is a tradeoff between image quality and radiation dose in CT scans, with high dose scans producing high quality images but exposing patients to more radiation, and low dose scans resulting in increased image noise, particularly with low kV CT scans.
Innovation Solution
A method and system using a probabilistic model and machine learning to synthesize high dose or high kV CT images from low dose or low kV images by calculating voxel values based on a likelihood function and prior function, allowing for the generation of high quality images with improved radiation dose or noise reduction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If high dose CT scans are used, then image quality is improved, but radiation exposure to patients increases
Solution Approach 1:
The patent creates a virtual high dose CT image as a copy of what a high dose scan would produce, without actually performing the high dose scan. The system uses a probabilistic model trained on paired low-dose and high-dose images to synthesize a high-dose quality image from a low-dose input, thereby copying the desired image quality while avoiding the harmful radiation exposure
Solution Approach 2:
The patent introduces a probabilistic model as an intermediary between the low-dose input image and the desired high-dose output image. This model acts as a mediator that transforms the low-quality input into high-quality output without requiring direct high-dose exposure, using learned statistical relationships from training data to bridge the gap between low and high dose images
2Object-affected harmful factors
If low kV CT scans are used, then radiation dose is reduced, but image noise increases
Solution Approach 1:
The patent synthesizes a virtual high kV image that copies the low-noise characteristics of high kV scans while maintaining the low radiation dose benefit of low kV acquisition. The probabilistic model learns the statistical relationship between low kV and high kV images during training, enabling it to generate a high-quality output that resembles what a high kV scan would produce without actually using high kV parameters
Solution Approach 2:
The patent changes the effective kV parameter of the output image through computational processing rather than physical adjustment. By using a probabilistic model trained on kV-paired images, the system transforms a low kV input into an image that has the statistical properties of a high kV image, effectively changing the kV parameter in the digital domain while maintaining the low radiation dose of the original low kV acquisition
Data Source
AI summary
A method and apparatus for medical image synthesis is disclosed, which synthesizes a target medical image based on a source medical image. The method can be used for synthesizing a high dose computed tomography (CT) image or a high kV CT image from a low dose CT image or a low kV image. A plurality of image patches are extracted from a source medical image. A synthesized target medical image is generated from the source medical image by calculating voxel values in the synthesized target medical image based on the image patches extracted from the source medical image using a machine learning based probabilistic model.