Tomosynthesis Image Processing Using Fuzzy Logic for Lesion Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current mammography methods using tomosynthesis struggle with efficiently detecting radiological signs due to high data volume and time-consuming information access, leading to potential overlooks of lesions and difficulties in analyzing small radiological signs within dense breast tissues.
Innovation Solution
A method and apparatus employing fuzzy logic for image processing that reconstructs 3D images from radiography projections, determines candidate particles, assigns membership degrees to classes, aggregates these particles to form 3D fuzzy particles, and calculates confidence levels for radiological signs, reducing the time required to locate these signs and improving diagnostic accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If digital tomosynthesis screening methods are used to produce 3D images, then the detection of lesions masked during superimposition is improved, but the quantity of information to be managed increases greatly and access time to clinically interesting information becomes very long
Solution Approach 1:
The patent extracts and isolates only the clinically relevant information from the large 3D dataset by using automated detection algorithms to identify and highlight suspicious areas. This extraction process removes unnecessary data from the radiologist's view, allowing quick access to clinically interesting information without manually searching through the entire volume.
Solution Approach 2:
The patent introduces an automated detection system as an intermediary between the 3D image data and the radiologist. This intermediary automatically processes the large dataset, identifies potential lesions, and presents filtered results to the radiologist, thereby reducing access time while maintaining reliable detection.
2Reliability
If tomosynthesis mammography methods are used to detect lesions, then the visibility of structures disturbed by superimposition is improved, but the frequency of use cannot be high due to long information access time
Solution Approach 1:
The patent performs preliminary processing of the 3D image data using automated detection algorithms before the radiologist views the images. This preliminary action identifies and flags suspicious areas in advance, so that when the radiologist accesses the images, only the relevant suspicious areas need attention, dramatically reducing access time and enabling higher screening frequency.
Solution Approach 2:
The patent replaces the manual mechanical search process with automated computational algorithms. Instead of the radiologist manually navigating through 50-80 slices to find lesions, an automated system rapidly processes and highlights suspicious areas, substituting computational speed for manual searching and enabling higher productivity.
3Measurement precision
If high resolution detection is used to identify small radiological signs between 100 pm and 1 mm, then the detection precision is improved, but the time required for rapid searches in large volumes increases
Solution Approach 1:
The patent applies local quality by using automated detection algorithms that focus computational resources only on local regions containing suspicious features. Instead of uniformly processing the entire large volume at high resolution, the system identifies regions of interest and applies detailed analysis only where needed, maintaining high detection precision for small objects while reducing overall search time.
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 significantly reduces the time to detect radiological signs, enhances diagnostic confidence by highlighting all signs in 3D images, and automatically distinguishes malignant from benign elements, thereby improving the effectiveness of mammography screenings.
Implementation Method 1
a source of X-ray radiation
Implementation Method 2
a detector of the X-ray radiation passing through an organ to be imaged
Data Source
AI summary
A method of image processing in a radiological apparatus includes reconstructing a 3D image of a body from a set of radiography projection images, locating structures presumed to be representative of 3D radiological signs within the 3D image, determining a set of 2D candidate particles corresponding to projections of the presumed 3D radiological signs, assigning, through a fuzzy logic description, to the 2D candidate particles a degree of membership in 2D membership classes of a set of membership classes, each membership class being relative to a type of radiological sign, considering a 2D fuzzy particle being formed by the set of the 2D candidate particles and by their respective degrees of membership in a class, making an aggregate of the 2D fuzzy particles to obtain 3D fuzzy particles in a digital volume, and determining a degree of confidence for each 3D radiological sign from the 3D fuzzy particles.