A topology constrained geometric deformable model for medical image segmentation. (February 2021)

• Record Type: Journal Article
• Authors: Yahia Lahssene, Yamina; Meddeber, Lila; Zouagui, Tarik; Jennane, Rachid
• Is Part Of: Biomedical signal processing and control. Volume 64(2021)
• Publication Date: 2021-02
• Subjects: Image segmentation -- Geometric deformable models -- Level sets -- Topology preservation -- Heart ventricles; Signal processing -- Periodicals; Biomedical engineering -- Periodicals; Signal Processing, Computer-Assisted -- Periodicals; Image Processing, Computer-Assisted -- Periodicals; Biomedical Engineering -- Periodicals; 610.28
• DOI: 10.1016/j.bspc.2020.102299
• Languages: English
• ISSNs: 1746-8094

Abstract

Highlights: Geometric deformable model for global or local medical image segmentation. Effective heart endocardium ventricle extraction using a topology controlled selective binary level set model. A topology constrained model for segmentation tested on original cardiac MRI of MICCAI and York datasets. Improved cardiac MRI segmentation performance compared to different existing models. Active contour model for segmentation. Abstract: Background: The analysis and segmentation of Magnetic Resonance images are usually performed by experts. This process is tedious and time consuming. While many automatic and semi-automatic methods have been proposed in the literature, most of them are not reliable or accurate enough for medical image segmentation. Geometric Deformable Models have been widely used as a method of medical image segmentation using only geometric measures. They have proved their efficiency, ease of implementation and ability to handle automatic topology changes. Since, in medical images, the topology of the target to be segmented is known beforehand, different concepts of topology preserving have been implemented along with Geometric Deformable Models as a term to constrain their topology while evolving. Method: In this paper, we propose a new approach for image segmentation, called the Selective Binary Level Set function and a new variant of the Topology Preserving Selective Binary Level Set model. Results: The proposed approach successfully achieves either global or local segmentations. Different metrics were used to evaluate and compare our segmentation results of the heart ventricle to other methods. Conclusion: Results show the effectiveness of our proposed method, which proved accurate as a primary step for medical image analysis.