Tomographic Image Generation Device Using Weighted Pixel Projection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current tomographic image generation techniques face challenges in reducing artifacts and noise while maintaining image quality and reducing calculation time, particularly in tomosynthesis imaging, where the limited angles of radiation application and low radiation doses lead to visible noise and artifacts in reconstructed images.
Innovation Solution
A tomographic image generation device and method that projects pixel values from projection images onto a desired slice plane based on the positional relationship between the radiation source and detector, using regression analysis to calculate pixel values at each coordinate position, incorporating interpolation and filtering to reduce noise and artifacts, and adjusting sampling intervals for improved resolution and sharpness.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If a low radiation dose is used for each imaging to reduce radiation exposure, then radiation exposure is reduced, but quantum noise in the projection images increases resulting in more visible noise in the reconstructed tomographic image
Solution Approach 1:
The patent combines multiple projection images obtained at different radiation source positions through cumulative addition. By merging the information from multiple low-dose images, the signal is reinforced while random noise tends to cancel out, producing a tomographic image with acceptable quality despite each individual image being noisy
Solution Approach 2:
The patent applies radiation continuously at multiple different radiation source positions rather than using a single high-dose exposure. This continuous imaging approach from multiple angles allows accumulation of useful signal information while distributing the total radiation dose across multiple lower-dose measurements
2Device complexity
If a reverse projection method is used to reconstruct the tomographic image, then the reconstruction process is simple, but artifacts appear as virtual images of structures in areas where they are not actually present
Solution Approach 1:
The patent introduces weighting factors as an intermediary element in the reconstruction process. These weighting factors, calculated based on similarity between projection images, mediate between the simple reverse projection method and the need to reduce artifacts, allowing the basic method to be enhanced without complete redesign
Solution Approach 2:
The patent modifies the reconstruction process by changing the parameter of pixel weighting. Instead of uniform weighting in simple reverse projection, the patent dynamically adjusts weighting factors based on image similarity metrics, transforming the reconstruction from a simple arithmetic operation to a weighted operation that suppresses artifacts
3Manufacturing precision
If weighting factors are calculated based on similarity between projection images to reduce artifacts, then artifact reduction is achieved, but calculation time increases significantly
Solution Approach 1:
The patent applies different weighting factors to different pixel positions based on local similarity characteristics. Rather than uniformly processing all pixels with equal computational effort, the method focuses computational resources on pixels where artifact reduction is most needed, improving efficiency by adapting the level of processing to local image characteristics
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach significantly reduces artifacts and noise, enhances image quality, and decreases calculation time by considering pixel values around the coordinate position of interest, allowing for higher resolution and sharper images while optimizing computational efficiency.
Implementation Method 1
a radiographic imaging apparatus using radiation, such as x-ray or γ-ray
Implementation Method 2
applying radiation to the subject from different radiation source positions
Data Source
AI summary
An image obtaining unit obtaining a plurality of projection images taken by imaging a subject with different radiation source positions. A pixel value projecting unit projects pixel values of the projection images on coordinate positions on a desired slice plane of the subject based on the positional relationship between the radiation source position with which each projection image is taken and the radiation detector, while preserving pixel values of the projection images. A pixel value calculating unit calculates a pixel value at each coordinate position of interest on the slice plane based on a plurality of pixel values of the projection images projected in a predetermined range relative to the coordinate position of interest on the slice plane to thereby generate a tomographic image of the slice plane.