Tomosynthesis Microcalcification Visibility via Frequency-Divided Slice Thickness
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Digital breast tomosynthesis methods struggle to clearly depict microcalcifications, especially in wide-angle recordings, due to limited visibility and susceptibility to intensity differences in thresholding methods, leading to brief 'flare-ups' when switching between slice images.
Innovation Solution
A method generating result slice images with varying slice thickness based on tomosynthesis data, using frequency division and mixing of average and maximum value slices to enhance microcalcification visibility, where slice thickness differs depending on spatial frequency, allowing microcalcifications to be depicted across multiple slices and improving contrast.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Duration of action of moving object
If a single uniform slice thickness is used in tomosynthesis, then the processing is simple, but microcalcifications are only visible in very few slice images resulting in brief flare-up
Solution Approach 1:
The patent applies local quality by using different slice thicknesses for different spatial frequencies. High-frequency components (microcalcifications) use increased slice thickness for extended visibility, while low-frequency components (morphological structures) use standard slice thickness. This is implemented through frequency-domain processing where the tomosynthesis data is divided into frequency components, and each component is reconstructed with its optimal slice thickness.
Solution Approach 2:
The patent segments the tomosynthesis data into different frequency components (low-pass and high-pass filtered components). This segmentation allows independent processing of each frequency type, enabling microcalcifications to be enhanced with increased slice thickness while maintaining morphological structures with standard thickness, thereby extending the visibility duration of microcalcifications across multiple slice images.
2Measurement precision
If thresholding method is used to obtain microcalcifications, then the process is simple, but the method is susceptible to intensity differences and does not always lead to optimal results
Solution Approach 1:
The patent changes the fundamental parameter of slice thickness based on spatial frequency rather than using fixed thresholding. By adjusting slice thickness as a function of frequency, the method becomes robust to intensity differences since it relies on structural frequency characteristics rather than intensity thresholds, thereby improving measurement precision without significantly increasing processing complexity.
Solution Approach 2:
The patent replaces the mechanical thresholding approach with a frequency-domain based approach. Instead of using intensity thresholding (a simple but unreliable method), the system uses frequency filtering and slice thickness variation, which provides more reliable microcalcification detection by leveraging the frequency characteristics of microcalcifications rather than their intensity.
3Measurement precision
If wide-angle recording is used, then the depth resolution is higher, but the depiction of microcalcifications is less clear compared to small-angle recording
Solution Approach 1:
The patent applies local quality by selectively enhancing high-frequency components (microcalcifications) with increased slice thickness while maintaining the full angular range data for morphological structures. This allows the system to retain the versatility of wide-angle recording while improving microcalcification clarity through frequency-specific processing.
Solution Approach 2:
The patent segments the tomosynthesis data into frequency components, allowing the high-pass filtered components (microcalcifications) to be processed with increased slice thickness while the low-pass components (morphological structures) use standard thickness. This segmentation enables simultaneous optimization for both microcalcification clarity and angular range coverage.
Data Source
AI summary
A method for generating result slice images with at least partially different slice thickness based on a tomosynthesis image data set of a breast includes generating average value slices and maximum value slices (MIP) based on the tomosynthesis image data set, frequency dividing the average value slices into low-pass filtered and high-pass filtered average value slices, high-pass filtering of maximum value slices to form high-pass filtered maximum value slices, mixing high-pass filtered maximum value slices and high-pass filtered average value slices to form mixed high-pass filtered maximum value slices, combining the low-pass filtered average value slices with the mixed high-pass filtered maximum value slices to form the result slice images, and applying a moving maximum value across a selected thickness of maximum value slices or across a selected thickness of mixed high-pass filtered maximum value slices.