Image Reconstruction via Sparse Transforms for Low-Dose CT
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In medical imaging, particularly in low-dose scenarios, existing technologies face challenges in maintaining image quality due to increased noise and decreased signal-to-noise ratio, leading to suboptimal reconstructed images.
Innovation Solution
A method involving multiple sparse transforms, such as gradient and wavelet transforms, followed by weighted reconstruction, is employed to enhance image boundary preservation and quality, aligning with Compressed Sensing theory to effectively reconstruct images from limited data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If radiation dose is reduced in low-dose imaging, then harmful radiation exposure is decreased, but image quality deteriorates due to increased noise and decreased signal-to-noise ratio
Solution Approach 1:
The patent applies preliminary action by performing multiple sparse transforms (gradient transform, wavelet transform, and second sparse transform) on the initially-updated image before final reconstruction. These preprocessing steps extract image boundary prior information and enhance edge details in advance, allowing the reconstruction algorithm to work with pre-enhanced data that compensates for the low signal-to-noise ratio caused by reduced radiation dose, thereby maintaining image quality while using lower radiation doses
2Manufacturing precision
If multiple sparse transforms are performed to enhance image boundaries, then image quality and boundary preservation are improved, but computational complexity increases
Solution Approach 1:
The patent applies segmentation by dividing the image processing into distinct sequential stages: gradient transform to extract edge information, wavelet transform to decompose the image into frequency components, and a second sparse transform to further enhance boundaries. Each transform focuses on specific image features, allowing the system to achieve superior boundary preservation through specialized processing steps rather than a single complex operation
Data Source
AI summary
Methods, devices and apparatus for reconstructing an image are provided. According to an example of the method, scanning data is obtained for a scanned subject, an initially-updated image is reconstructed from the scanning data, image boundary prior information is generated by performing at least two sparse transforms on the initially-updated image, and a reconstructed image is obtained by performing a weighted reconstruction with the image boundary prior information and the initially-updated image.