What is it about?
Access our research at: https://annals-csis.org/Volume_2/pliks/204.pdf STUDY AIMS The main aim of this paper is to advance the state of the art in automated prostate segmentation using T2weighted MR images, by introducing a hybrid topological MRI prostate segmentation method which is based on a set of pre-labeled MR atlas images. The proposed method has been experimentally tested on a set of 30 MRI T2 weighted images. For evaluation the automated segmentations of the proposed scheme have been compared with the manual segmentations, using an average Dice Similarity Coefficient (DSC). Obtained quantitative results have shown a good approximation of the segmented prostate.
Featured Image
Photo by National Cancer Institute on Unsplash
Why is it important?
- Completely new algorithm for MRI segmentation proposed. - Very easy to understand and implement. - Obtained results have shown a good approximation of the segmented prostate.
Perspectives
Comparing the obtained results a conclusion for prostate shape accordance can be derived, based on the prostate edges’ direction compatibility, prostate contour position and prostate surface. In general, prostate segmentation result in this case depends of two factors. The number of segmented samples used and segmented samples’ prostate shape variability, based on what the interception and the union shapes are determined. More segmented samples are considered, with wider prostate shape variability, more accurate prostate contour is obtained. A drawback of the proposed method is the incapacity of detecting prostate segments, out of the derived union region, being part of prostate of a non-segmented sample. On the opposite, prostate segmentation running time is significantly improved, since relatively small segment of a non-segmented prostate MR image is processed.
Ph.D Done Stojanov
Univerzitet Goce Delcev Stip
Read the Original
This page is a summary of: Topological Prostate Segmentation Method in MRI, September 2014, Polish Information Processing Society PTI,
DOI: 10.15439/2014f204.
You can read the full text:
Contributors
The following have contributed to this page
