Fracture Detection in Digitized X-Rays via Pixel Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
There is a high incidence of missed fractures in digitized x-rays, particularly minimally displaced or 'hairline cracks,' which can lead to devastating medical and legal consequences due to delayed or inappropriate treatment.
Innovation Solution
A method and apparatus for analyzing digitized x-ray images by representing them as a two-dimensional matrix of pixels, performing image processing, and using machine learning techniques to identify bone segments with a high likelihood of fracture, flagging potential fractures for further evaluation, and reducing false indications through curvature analysis and edge detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If image processing algorithms are used to detect fractures, then detection sensitivity is improved, but false indications increase
Solution Approach 1:
The patent divides the x-ray image into multiple regions of interest (ROIs) corresponding to different bones, and further segments each bone into multiple segments along its longitudinal axis. This segmentation allows the system to analyze specific regions independently, improving detection precision while reducing false positives by focusing analysis on anatomically relevant areas rather than the entire image.
Solution Approach 2:
The patent applies different evaluation criteria and thresholds to different bone segments based on their specific anatomical characteristics. Each bone segment is evaluated using parameters tailored to its location and morphology, allowing the system to adapt to local variations in bone structure and reduce false indications while maintaining high detection sensitivity.
2Measurement precision
If machine learning techniques are applied to identify bone segments, then detection accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent performs preliminary processing steps including bone segmentation and region of interest identification before applying the machine learning fracture detection algorithm. By pre-organizing the image data into structured bone segments with associated anatomical information, the system reduces the computational burden during the actual fracture detection phase while maintaining high accuracy.
Solution Approach 2:
The patent segments the x-ray image into distinct bone regions and further divides each bone into multiple segments, allowing the machine learning model to process smaller, more manageable data portions. This segmentation strategy reduces computational complexity by limiting the model's analysis to specific anatomical regions rather than the entire image.
3Measurement precision
If pixel intensity evaluation is performed to detect black pixels, then fracture detection capability is improved, but processing time increases
Solution Approach 1:
The patent focuses the pixel intensity evaluation on specific regions of interest corresponding to bone segments rather than analyzing the entire x-ray image. By limiting the evaluation to anatomically relevant areas where fractures are most likely to occur, the system maintains high detection capability while significantly reducing processing time.
Solution Approach 2:
The patent divides the image into multiple bone segments and evaluates pixel intensities within each segment independently. This segmentation allows the system to process only the necessary regions containing potential fractures, reducing the total number of pixels that require analysis and thereby decreasing overall processing time while maintaining detection precision.
Data Source
AI summary
A method for identifying one or more fractures in a digitized x-ray image includes: obtaining a digitized x-ray image; performing preprocessing on the digitized x-ray image to generate a modified x-ray image having enhanced resolution; partitioning the modified x-ray image into a two-dimensional array including multiple pixels; obtaining an intensity of each of the pixels in the modified x-ray image; evaluating each of the pixels in the modified x-ray image to determine whether at least a given one of the pixels has an intensity indicative of a black pixel; for each given pixel having an intensity indicative of a black pixel, flagging the given pixel when pixels immediately adjacent to the given pixel have an intensity greater than a prescribed threshold value; and placing a visual indication of a possible fracture on the modified x-ray image corresponding to a location of each flagged pixel.