Medical Image Boundary Correction for Attached Abnormal Shadows
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current computer-aided diagnosis (CAD) technologies struggle to accurately calculate shape feature amounts for attached abnormal shadows in medical images, as existing methods fail to account for the differences between attached and non-attached surfaces, leading to inaccurate determination of the center of gravity and inability to calculate features like spherical harmonics and ellipsoid-based metrics.
Innovation Solution
An image processing apparatus that extracts and corrects the shape of attached boundaries in medical images by identifying and differentiating between attached and non-attached surfaces, using techniques such as boundary extraction, identification, and correction units to refine the shape features for improved CAD analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the shape feature amount is calculated including the attached surface of an attached abnormal shadow, then the calculation can be performed using the entire boundary, but the calculated shape feature amount differs from the shape feature amount of the non-attached surface that the doctor wants to observe
Solution Approach 1:
The boundary of the first site is divided into a first boundary part (attached surface) and a second boundary part (non-attached surface). The correction unit uses the second boundary part to correct the shape of the first boundary part, enabling accurate calculation of shape feature amounts that reflect the non-attached surface characteristics while maintaining calculation efficiency.
2Ease of manufacture
If the center of gravity is determined using the region of the non-attached abnormal shadow, then the calculation is simple, but the center of gravity of the region of the attached abnormal shadow cannot be determined accurately
Solution Approach 1:
The second boundary part (non-attached surface) serves as an intermediary to correct the shape of the first boundary part (attached surface). This intermediary approach allows the center of gravity to be accurately determined for the attached abnormal shadow region by using the reliable non-attached surface geometry as a reference, without requiring complex direct measurement of the attached region.
3Ease of operation
If existing CAD methods are used for attached abnormal shadows, then the processing can be performed without special correction, but shape feature amounts such as spherical harmonics and ellipsoid-based metrics cannot be calculated accurately
Solution Approach 1:
The correction unit performs preliminary correction of the first boundary part shape using the second boundary part before shape feature amount calculation. This preliminary action ensures that subsequent calculations of spherical harmonics, ellipsoid-based metrics, and other shape features are performed on corrected geometry that accurately represents the non-attached surface, thereby improving CAD diagnostic accuracy while maintaining operational convenience.
Data Source
AI summary
A boundary of a site included as an attached abnormal shadow in a medical image is corrected to allow assisting in diagnosis based on a shape of a tissue. An image processing apparatus according to the present invention includes: a boundary extraction unit configured to extract a boundary of a first site and a boundary of a second site in the medical image; a boundary identification unit configured to identify a partial boundary of the first site that is adhere to the boundary of the second site as a first boundary part, and to identify a partial boundary of the first site that is not adhere to the boundary of the second site as a second boundary part; and a correction unit configured to correct a shape of the first boundary part based on the second boundary part.