Mass Density Candidate Selection in Digital Mammograms
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for selecting mass density candidates from mammograms rely heavily on intensity-based algorithms, which require numerous parameter adjustments and result in expensive computations, especially when dealing with continuous intensity levels and sizes, and are inefficient in processing time.
Innovation Solution
A method that combines intensity and morphologic algorithms, down-sampling images, applying a Gaussian difference filter, masking with a breast mask, and using a Canny detector to find mass density contours, processing each image at a single gray-level to enhance sensitivity and speed.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If intensity-based algorithms are used to select mass density candidates, then detection sensitivity is improved, but computation cost and processing time increase significantly
Solution Approach 1:
The patent segments the detection process into two distinct phases: a fast preprocessing phase using morphological operations to generate initial candidate regions, and a subsequent refinement phase that applies intensity-based analysis only to these candidates. This segmentation allows the system to maintain high detection sensitivity while dramatically reducing overall processing time by avoiding exhaustive intensity analysis of the entire image.
Solution Approach 2:
The patent performs preliminary action by first applying morphological operations (opening, closing, and top-hat transforms) to the mammogram image to pre-identify potential mass density regions before intensity-based analysis. This preliminary processing creates a reduced set of candidate regions that are then subjected to more computationally intensive intensity analysis, thereby improving processing efficiency while maintaining detection sensitivity.
2Measurement precision
If multiple Fourier band-pass images are processed to select mass densities, then detection accuracy is improved, but processing time increases to over 5 seconds
Solution Approach 1:
The patent extracts and applies only the essential morphological operations (opening, closing, top-hat transforms) that provide the most significant contribution to mass density detection, rather than processing multiple Fourier band-pass images. This extraction of key operations maintains detection accuracy while reducing processing time from over 5 seconds to under 1 second.
Solution Approach 2:
The patent changes the processing parameters by operating on the original image space using morphological operations with specifically chosen structuring elements, rather than transforming to frequency space with multiple Fourier bands. This parameter change in the processing domain achieves similar detection accuracy with significantly reduced computational complexity and processing time.
3Adaptability or versatility
If ad hoc adjustment of numerous parameters is performed to handle continuous intensity levels, then detection flexibility is improved, but device complexity and ease of operation deteriorate
Solution Approach 1:
The patent implements self-service through automatic parameter selection in the morphological operations. The structuring element sizes and operation parameters are determined algorithmically based on image characteristics rather than requiring manual ad hoc adjustment. This automation maintains detection flexibility across different mammograms while eliminating the complexity of manual parameter tuning.
Solution Approach 2:
The patent employs parameter changes by using adaptive morphological operations where structuring element sizes and transformation parameters are automatically adjusted based on local image characteristics. This automatic parameter adaptation provides detection flexibility for various mass densities and sizes without requiring manual intervention or complex parameter adjustment interfaces.
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 improves sensitivity and significantly reduces processing time, generating mass density candidates in less than 500 ms compared to traditional band-pass methods, which take over 5 seconds, and can be applied to various imaging modalities including ultrasound and MRI.
Implementation Method 1
applying a Gaussian difference filter to enhance intensity to form a filtered image
Implementation Method 2
using a Canny detector to find potential mass density contours
Data Source
AI summary
The present invention provides a method for selecting mass density candidates from digital image, for example mammograms, for computer-aided lesion detection, review and diagnosis. A method of selecting mass density candidates from a digital image for computer-aided cancer detection, review and diagnosis includes down-sampling the digital image to a low resolution; smoothing an edge along a skinline; applying a Gaussian difference filter to enhance intensity to form a filtered image; masking the filtered image using a breast mask; using a Canny detector to find potential mass density contours; and generating a mass density candidate list from Canny contours produced in the Canny detector.