Tomosynthesis Image Processing Using Mexican Hat Wavelet Filters
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current mammography devices face challenges in accurately detecting radiological signs due to superimposed structures in 2D projections, leading to falsely positive or negative interpretations, and the high volume of data in 3D tomosynthesis images makes it time-consuming to access clinical information, limiting the frequency of use and requiring radiologists to search for small objects in large volumes.
Innovation Solution
An image-processing method using linear differential filters, specifically Mexican hat type wavelet filters, to enhance and identify radiological signs in 3D data by computing contrast and validating voxels against predefined conditions, reducing the time to locate these signs and improving data readability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If 3D tomosynthesis imaging is used to improve detection accuracy, then the ability to detect radiological signs is improved, but the time required to access and review clinical information increases significantly
Solution Approach 1:
The patent segments the 3D breast volume into multiple axial slices, allowing the radiologist to review the breast anatomy in a systematic layer-by-layer manner. This segmentation enables efficient navigation through the large volume of 3D data by dividing it into manageable sections, reducing the time required to access and review clinical information while maintaining detection accuracy.
Solution Approach 2:
The patent applies automatic detection algorithms that pre-process the 3D data to identify and highlight potential radiological signs before the radiologist performs manual review. This preliminary action filters and prioritizes the data, reducing the time needed for the radiologist to access clinically relevant information by pre-organizing and flagging potential findings.
2Productivity
If standard 2D mammography projections are used, then the examination process is fast, but superimposed structures cause falsely positive or negative interpretations
Solution Approach 1:
The patent transitions from 2D projection imaging to 3D volumetric tomosynthesis imaging, adding the depth dimension to the analysis. This dimensional change allows the radiologist to view structures in multiple planes and depths, separating superimposed structures that appear together in 2D projections. The 3D visualization enables accurate differentiation between overlapping tissues, eliminating falsely positive and negative interpretations while maintaining examination efficiency through automated processing.
3Measurement precision
If high-resolution detection is applied to find small objects in large volumes, then the detection of radiological signs between 100 μm and 1 mm is improved, but the search time increases significantly
Solution Approach 1:
The patent introduces automatic detection algorithms as an intermediary between the raw 3D data and the radiologist's analysis. These algorithms act as a mediator that processes the large volume of high-resolution data, identifies potential radiological signs, and presents them to the radiologist for confirmation. This intermediary system handles the time-consuming task of searching through large volumes for small objects, allowing the radiologist to focus on verification and diagnosis.
Solution Approach 2:
The patent employs contrast enhancement and color-coding techniques to highlight radiological signs within the 3D volume. By applying different color intensities or contrast levels to regions with potential abnormalities, the system makes small objects (100 μm to 1 mm) visually distinguishable from surrounding tissues. This visual enhancement reduces search time by directing the radiologist's attention to potential findings rather than requiring manual scanning of the entire volume.
Data Source
AI summary
In an image-processing method for the detection of radiological signs in series of 3D data, an algorithm is used to detect radiological signs in a digital volume according to their contrasts. This algorithm is applied to reconstructed slices or directly to the series of projections. This algorithm is made by means of linear differential filters for signal analysis. It is used to color or enhance the intensity of the detected radiological signals according to the degree of malignancy.