Adaptively Weighted Anisotropic Diffusion for CT Image Denoising
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing denoising techniques in computed tomography (CT) imaging struggle to effectively reduce noise while preserving clinically significant structures, particularly small structures, as they often smooth edges and reduce image resolution, leading to loss of clinically relevant details.
Innovation Solution
The implementation of adaptively weighted anisotropic diffusion (AWAD) in post-reconstruction processing, which incorporates a time-variant and spatially-variant adaptive weight term to adjust the diffusion process at each iteration, enhancing edge preservation and noise reduction around both large and small structures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If low-pass filtering techniques are used to reduce noise, then noise is reduced and detectability of large objects is improved, but clinically significant structures and small structures are lost due to edge smoothing
Solution Approach 1:
The patent applies different filtering strengths to different regions of the image based on local structure characteristics. The anisotropic diffusion filter adapts its smoothing behavior locally, applying stronger filtering in homogeneous regions while preserving edges and fine structures where gradients are detected, thus resolving the contradiction between noise reduction and structure preservation
Solution Approach 2:
The patent employs an iterative diffusion process where the filter dynamically adjusts its behavior based on local image gradients and noise estimates at each iteration. This dynamic adaptation allows the filter to progressively reduce noise while maintaining structural integrity, overcoming the static limitations of traditional low-pass filters
2Manufacturing precision
If anisotropic diffusion is used to preserve certain predetermined structures, then those specific structures are improved, but general clinically significant structures are not preserved
Solution Approach 1:
The patent extends anisotropic diffusion to handle multiple types of clinically significant structures simultaneously by using adaptive weighting based on local noise characteristics and gradient magnitude. The filter becomes universal in its ability to preserve various structure types (vessels, lesions, edges) rather than being specialized for predetermined structures only
3Manufacturing precision
If prior art anisotropic smoothing techniques are used to preserve small structures, then small structures are improved, but the method requires determining an importance map which increases complexity
Solution Approach 1:
The patent enables the filter to automatically adapt to different structures without requiring manual importance map creation. The adaptive anisotropic diffusion filter self-adjusts its parameters based on local image characteristics, eliminating the need for external importance maps while maintaining small structure preservation capabilities
Data Source
AI summary
Embodiments and processes of computer tomography perform tasks associated with denoising a reconstructed image using an anisotropic diffusion filter and adaptively weighting an iterative instance of the diffused image based upon the product of a weight value and a difference between the iterative instance of the diffused image and the original image. In general, the adaptive weighting is a negative feedback in the iterative steps.