Medical Image Misregistration Correction via Iterative Reprojection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current medical imaging techniques, such as PET and SPECT, face challenges with attenuation artifacts due to misregistration of transmission and emission datasets, leading to inaccurate attenuation correction and increased radiation exposure to reduce image noise, which can result in misdiagnoses and loss of image contrast.
Innovation Solution
The system employs iterative reprojection of reconstructed datasets and a genetic artificial intelligence technique to correct misregistration and reduce noise in images, creating a more representative distribution of image activity without sacrificing contrast, using a computer-implemented method that includes creating correction masks and filtering to enhance image data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If image filters are applied to reduce noise, then image noise is reduced, but image contrast is lost
Solution Approach 1:
The patent applies different filtering operations to different regions of the image based on local characteristics. The system identifies edges and boundaries where contrast must be preserved, while applying stronger noise reduction in homogeneous regions. This local differentiation allows noise reduction without sacrificing the contrast needed for diagnostic accuracy.
Solution Approach 2:
The filtering approach is dynamic rather than static. The system adaptively adjusts filtering strength based on local image features, radiation dose levels, and noise characteristics. This dynamic adaptation allows the filter to maintain contrast in critical regions while effectively reducing noise in other areas, resolving the contradiction between noise reduction and contrast preservation.
2Reliability
If radiation dose is increased to reduce image noise, then image noise is reduced, but patient radiation exposure increases
Solution Approach 1:
The patent replaces the mechanical approach of increasing radiation dose with an intelligent software-based noise reduction system. Instead of physically adding more radiation to improve signal quality, the system uses advanced algorithms to distinguish between noise and actual image features, achieving noise reduction without additional radiation exposure to the patient.
Solution Approach 2:
The system changes the parameters of image processing rather than physical imaging parameters. By adjusting filtering strength, threshold values, and processing algorithms, the system achieves equivalent noise reduction效果 that would traditionally require increasing radiation dose, thereby protecting patients from unnecessary exposure while maintaining diagnostic image quality.
3Measurement precision
If misregistration correction is performed prior to emission reconstruction, then attenuation correction accuracy is improved, but raw sinogram data must be available
Solution Approach 1:
The patent performs misregistration correction as a preliminary step before emission reconstruction, but does so in a way that creates a reusable correction map. This preliminary correction establishes accurate alignment between transmission and emission data, and the resulting correction parameters are stored and applied subsequently, eliminating the need to repeatedly access raw sinogram data while maintaining correction accuracy.
Solution Approach 2:
The system creates a digital copy or representation of the misregistration correction in the form of a correction map or transformation parameters. Instead of requiring the original raw sinogram data to be permanently stored and accessible, the correction is captured in a compressed parameter form that can be applied later without needing the original complex datasets, reducing data availability requirements while preserving correction precision.
Data Source
AI summary
A system and computer-implemented method for improving the quality of images, such as images used in medical diagnosis. In a first technique, sinogram data, intermediate volumes, and transmission CT-based attenuation correction files are created. Scatter correction is performed, and a scatter correction map is created. Misregistration offsets are measured, and a correction registration mask is created using volume reprojection. The two masks are combined and applied to create the enhanced image data. In a second technique, a difference may be determined for each parent and child pixel pair, and an average difference may be determined for all pairs. For each pair which is under the average difference, the child may be eliminated and replaced with a new child. This process may be iteratively repeated, with the image data being incrementally enhanced with each iteration. For both techniques, the enhanced image data may be communicated to an interpretive application for display.