Dilation Mechanism for Image Segmentation Edge Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing digital image segmentation techniques using region growing mechanisms often result in edge insets when employing a test area defined by a radius, leading to incomplete segmentation of objects, especially in medical imaging where structures like the vascular system present challenges due to narrow connections.
Innovation Solution
A dilation mechanism is implemented to perform post-processing on regions or volumes generated by test area-based region growing, expanding the edges to the proper boundaries within the tolerance range, correcting for edge insets and preventing growth down undesired pathways.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a test area defined by a radius is used in region growing, then the region growing mechanism can prevent growth down narrow pathways, but edge insets are created leading to incomplete segmentation
Solution Approach 1:
The patent applies preliminary action by performing a dilation operation after region growing to pre-compensate for the edge inset effect. The dilation mechanism expands the segmented region back toward the true edge, correcting the incomplete segmentation caused by the test area radius constraint during the growth process.
Solution Approach 2:
The patent changes the parameter approach by introducing a separate dilation radius parameter that is larger than the original test area radius. This parameter change allows the system to first grow regions conservatively using the smaller test area radius, then expand them using the larger dilation radius to achieve accurate edge detection.
2Measurement precision
If a larger test area radius is used, then edge insets are reduced, but the region growing mechanism may grow down undesired narrow pathways
Solution Approach 1:
The patent segments the image processing into two distinct phases: a first phase using a smaller test area radius to grow regions while preventing descent into narrow pathways, and a second phase using a larger dilation radius to expand the regions to their true boundaries. This segmentation of the processing steps allows each phase to optimize for its specific purpose.
Solution Approach 2:
The patent uses preliminary action by completing the cautious region growing first with the smaller radius, establishing a reliable initial segmentation that avoids undesired pathways, then applying the dilation operation as a subsequent corrective step to achieve accurate edges without risking pathway infiltration.
3Reliability
If a smaller test area radius is used, then growth down undesired pathways is prevented, but edge insets are created resulting in incomplete segmentation
Solution Approach 1:
The patent ensures continuity of useful action by chaining two operations together: the region growing operation that reliably prevents pathway descent, followed immediately by the dilation operation that continuously expands the segmented regions to their complete boundaries, maintaining productive segmentation throughout the process.
Solution Approach 2:
The patent applies preliminary action by first establishing a reliable region segmentation using the smaller test area radius that prevents undesired pathway growth, then using the dilation operation to pre-compensate for the edge inset and achieve complete segmentation, ensuring both reliability and productivity are maintained.
Data Source
AI summary
Method and apparatus for extending regions in two-dimensional (2-D) image space or volumes in three-dimensional (3-D) image space that are generated by a test area-based region growing mechanism. Embodiments of a dilation mechanism may perform post-processing of a region or volume generated by the test area-based region growing mechanism to correct for an edge inset resulting from a radius used to define the test area. The dilation mechanism may perform a morphological dilate to expand the region or volume to the proper edge of the desired object in the image data within the tolerance range of the threshold, and thus corrects for the inset error introduced by the test area radius used by the region growing mechanism. The dilation mechanism may be limited to extending the region or volume to the radius distance from the edge of the original region or volume generated by the region growing mechanism.