Facial reconstruction

Search LJMU Research Online

Browse Repository | Browse E-Theses

Fully Automatic Ultrasound Fetal Heart Image Detection and Segmentation based on Texture Analysis

Song, C, Gao, T, Gong, Y, Sudirman, S, Wang, H and Zhu, H (2020) Fully Automatic Ultrasound Fetal Heart Image Detection and Segmentation based on Texture Analysis. Investigacion Clinica, 61 (2). pp. 600-608. ISSN 0535-5133

[img]
Preview
Text
Fully Automatic Ultrasound Fetal Heart Image Detection and Segmentation based on Texture Analysis.pdf - Published Version

Download (1MB) | Preview

Abstract

Ultrasound fetal heart image analysis is important for the antenatal diagnosis of congenital heart disease, therefore, design an automated fetal heart ultrasound image analysis approaches to improve detection ratio of congenital heart disease is necessary. Nevertheless, because of the complicated structure of fetal heart ultrasound image, location, detection and segmentation approaches of fetal heart images as interesting topics that get more attention. Therefore, in this work, we present a framework to segment ultrasound image automatically for tracking the boundary of fetal heart region. In the first step, this paper contributes to breed candidate regions. And then, in the segmentation progress, we apply an energy-based active contour model to detect the edges of fetal heart. Finally, in the experiment section, the performance is estimated by the Dice similarity coefficient, which calculate the spatial overlap between two different segmentation regions, and the experiment results indicate that the proposed algorithm achieves high levels of accuracy.

Item Type: Article
Subjects: Q Science > QA Mathematics > QA75 Electronic computers. Computer science
R Medicine > R Medicine (General)
R Medicine > RG Gynecology and obstetrics
Divisions: Computer Science & Mathematics
Publisher: Facultad de Medicina de la Universidad del Zulia
Date Deposited: 24 Apr 2020 09:43
Last Modified: 24 Apr 2020 09:45
URI: https://researchonline.ljmu.ac.uk/id/eprint/12813

Actions (login required)

View Item View Item