X-ray Image Polluting Region Masking via Statistical Texture Synthesis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In medical radiography, particularly mammography, the presence of 'polluting regions' in X-ray images hampers radiologists' ability to detect anomalies, as existing image processing methods either fail to effectively isolate these regions or interfere with the analysis of regions of interest.
Innovation Solution
A computer system and method that identifies polluting regions in X-ray images by analyzing initial intensity values and metadata, replacing these regions with texturing elements that match the surrounding tissue type and features, thereby masking them without altering the image's natural appearance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If polluting regions are modified by applying identical intensity values or increasing intensity of regions of interest, then the polluting regions are suppressed, but the radiologist's analysis is interfered with and interpretation is hampered
Solution Approach 1:
The patent creates a copy of the polluting region and replaces it with a synthesized version generated from texturing elements. The synthesis process copies statistical features (mean, variance, skewness, kurtosis) from the original region and applies them to randomly selected texturing elements, creating a realistic-looking replacement that suppresses the polluting effect while preserving natural appearance for radiologist analysis
Solution Approach 2:
The patent transforms the polluting region by changing its intensity parameters. Instead of simple intensity equalization, it synthesizes new intensity values by selecting texturing elements and adjusting their parameters (mean, variance, skewness, kurtosis) to match the statistical characteristics of the original polluting region, thereby maintaining visual realism while reducing diagnostic interference
2Object-affected harmful factors
If image processing methods are applied to modify polluting regions, then the visibility of polluting regions is reduced, but the natural appearance of the image is altered
Solution Approach 1:
The patent changes the parameters of texturing elements (mean, variance, skewness, kurtosis) to match the statistical characteristics of the original polluting region. This parameter matching ensures that the synthesized replacement maintains the natural appearance and texture characteristics of the original image, making the modification imperceptible to radiologists
Solution Approach 2:
The patent copies the statistical distribution characteristics (mean, variance, skewness, kurtosis) from the original polluting region and applies them to synthesized texturing elements. This copying of statistical features ensures that the replacement region maintains the natural appearance and texture of the original image while suppressing the polluting effect
Data Source
AI summary
The invention relates to a system and method for processing a polluting region of a digital image of an object (3) acquired using X-rays and comprising points (P) with each of which there is associated a numerical intensity value and at least one item of metadata (M) chosen from among one or more physical parameters of the object, and one or more parameters relating to the power or the energy of the X-rays with which the image was acquired. Said method comprises identifying at least one group of points in the image corresponding to a polluting region on the basis of the initial intensity values I0 of the points in the image and of said metadata. Generating a new image comprises a step of replacing the initial intensity values I0 of at least some of the points of said at least one polluting region with new intensity values.