Multi-threshold Peripheral Equalization for Digital Mammography
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In mammography and breast tomosynthesis images, the peripheral areas often have lower intensity than the central areas due to varying breast thickness, requiring radiologists to adjust window levels, leading to increased reading time and undesirable artifacts like segmentation lines when using prior PE methods.
Innovation Solution
A peripheral equalization method using multiple thresholds to process pixel data, with a fixed intensity for the central part and variable thresholds for the peripheral part, eliminating segmentation lines and enhancing image visibility across the entire breast at a single window level.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Illumination intensity
If peripheral equalization methods are used to enhance peripheral area intensity, then image visibility is improved, but segmentation lines and overshoot banding artifacts are introduced
Solution Approach 1:
The patent applies different processing thresholds to different regions of the image. The peripheral area uses multiple thresholds to enhance intensity, while the central area uses a fixed threshold. This local differentiation allows peripheral equalization without introducing artifacts across the entire image, as each region is processed according to its specific characteristics.
Solution Approach 2:
The patent segments the breast image into central and peripheral areas based on spatial location. By dividing the image into these distinct regions, the system can apply different processing methods to each segment - multiple thresholds for the periphery and fixed threshold for the center - thereby achieving peripheral equalization without propagating artifacts to other regions.
2Illumination intensity
If window level adjustment is performed for different breast regions, then image quality is improved, but reading time increases
Solution Approach 1:
The patent performs preliminary processing by automatically equalizing the peripheral area intensity before the radiologist views the image. This pre-processing step prepares the image in advance, eliminating the need for the reader to manually adjust window levels for different regions, thereby reducing reading time while maintaining image quality.
Solution Approach 2:
The system performs self-service by automatically detecting and correcting the intensity imbalance between central and peripheral areas. The peripheral equalization algorithm autonomously processes the image without requiring user intervention, thereby saving the radiologist's time that would otherwise be spent manually adjusting window levels.
3Object-generated harmful factors
If multiple thresholds are used to process peripheral pixel data, then peripheral equalization is achieved without segmentation lines, but processing complexity increases
Solution Approach 1:
The patent applies multiple thresholds only to the peripheral area processing, while using a single fixed threshold for the central area. This localized application of complex processing only where needed reduces the overall processing complexity burden, as the computationally intensive multi-threshold operation is confined to a specific region rather than the entire image.
Solution Approach 2:
By segmenting the image into central and peripheral regions, the patent confines the complex multi-threshold processing to only the peripheral segment. This segmentation strategy isolates the computational complexity to a specific area, making the overall system more manageable and efficient compared to applying complex processing uniformly across the entire image.
Data Source
AI summary
A peripheral equalization (PE) method and apparatus for compensating for thickness reduction in outer edges of the breast in a mammogram (i.e. a two-dimensional image) while keeping the central area substantially unchanged. The PE method and apparatus can also be applied to three dimensional (tomosynthesis) images of a breast. The peripheral equalization is achieved by segmenting the image of the breast into at least two regions and using a multi-threshold technique to process the data in at least one of the two regions.