DBT Artifact Reduction via Statistical Outlier Rejection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Digital breast tomosynthesis and computer tomography face challenges in reducing artifacts caused by dense tissue and large masses, which complicate three-dimensional visualization and reconstruction, especially due to limited angular range acquisition, leading to out-of-plane blur and distortion of calcifications.
Innovation Solution
A method involving the generation of a statistical model based on pre-computed image data from multiple patients and phantoms to predict and remove outliers by comparing grey values with a stored model, ensuring accurate reconstruction and reducing artifacts without blurring small anatomical details.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If a slice thickness filter is used to suppress out-of-plane artifacts, then artifact reduction is improved, but small anatomical details such as micro calcifications are blurred
Solution Approach 1:
The patent applies different processing treatments to different regions of the image based on local characteristics. Dense tissue regions undergo outlier rejection processing to remove artifacts, while regions containing small anatomical details preserve their original information to maintain sharpness and visibility of micro calcifications.
Solution Approach 2:
The patent segments the image data into different regions based on tissue density and artifact presence. By identifying and separating dense tissue regions from other areas, the system can apply artifact reduction algorithms selectively only where needed, preserving small anatomical details in non-dense regions.
2Object-affected harmful factors
If statistical outlier rejection methods are used to remove artifacts, then some out-of-plane blur is reduced, but not all blur caused by large dense structures is removed because the limited angular range causes correct intensity information to be removed as outliers
Solution Approach 1:
The patent performs preliminary classification of projection images to identify those containing dense tissue structures before applying outlier rejection. By pre-identifying dense tissue regions and their corresponding projections, the system can preserve correct intensity information from these regions while still removing artifacts from other areas.
Solution Approach 2:
The patent dynamically adjusts the statistical parameters and thresholds for outlier detection based on the specific characteristics of each projection image and region. By adapting the outlier criteria to local image properties rather than applying uniform thresholds, the system preserves correct intensity information while removing artifacts.
3Object-affected harmful factors
If conventional CT artifact reduction algorithms are applied to tomosynthesis, then some artifacts are reduced, but they are not ideally suited for tomosynthesis images caused by dense tissue, metal clips and calcifications due to limited acquisition angle
Solution Approach 1:
The patent modifies the artifact reduction algorithm parameters specifically for tomosynthesis applications, adjusting the angular range considerations and statistical thresholds to account for the limited acquisition angle. This adaptation makes the algorithm effectively suited for tomosynthesis images with dense tissue, metal clips, and calcifications.
Solution Approach 2:
The patent introduces an intermediate classification step that identifies the type of artifact-causing structure (dense tissue, metal clip, or calcification) before applying the appropriate reduction strategy. This intermediary classification enables the system to select and apply the most effective artifact reduction method for each specific case.
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 effectively reduces artifacts, improves the visibility of clinical features, and maintains high contrast and sharpness in reconstructed images, enabling accurate three-dimensional visualization and reconstruction without the loss of detail.
Implementation Method 1
emitting X-ray radiation through a matter to be analyzed
Data Source
AI summary
Reduction of artifacts in digital breast tomosynthesis and in computed tomography. Because of the limited angular range acquisition in DBT the reconstructed slices have reduced resolution in z-direction and are affected by artifacts. Out-of-plane blur caused by dense tissue and large masses complicates 3D visualization and reconstruction of thick slices volumes. The streak-like out-of-plane artifacts caused by calcifications and metal clips distort the true shape of calcification, an important malignancy predictor. Microcalcifications could be obscured by bright artifacts. The technique involves reconstructing a set of super resolution slices and predicting the “artifact-free” voxel intensity based on the corresponding set of projection pixels using a statistical model learned from a set of training data. The resulting reconstructed images are de-blurred and streak artifacts are reduced, visibility of clinical features, contrast and sharpness are improved, 3D visualization and thick-slice reconstruction is possible without the loss of contrast and sharpness.