Patch-Based Fetal Heart Chamber Segmentation in Ultrasound Sequences Using Possibilistic Clustering

• Main Authors: Chisanai Sardsud, Sansanee Auephanwiriyakul, Nipon Theera-Umpon, Theera Tongsong
• Format: Conference Proceeding
• Published: 2018
• Subjects: Computer Science; Mathematics
• Online Access: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84994223724&origin=inward; http://cmuir.cmu.ac.th/jspui/handle/6653943832/55499
• Institution: Chiang Mai University

© 2015 IEEE. Fetus cardiomegaly, an abnormal enlargement of heart, is one of the early indicators of hemoglobin Bart's disease that is the most dangerous thalassemia in Thai fetuses. The cardiothoracic ratio or CT ratio measurement helps detecting the fetus cardiomegaly. An automatic measuring system is a helpful tool for the physicians. However, in order to do that we need to have an automatic heart chamber segmentation system. In this paper, we developed an automatic fetal heart chamber segmentation system using a patch-based possibilistic c-means. We extracted the region containing heart chambers and ultrasound video frames with the end diastolic and end systolic using the Horn-Schunck's optical flow algorithm. We also extracted the region containing the fetal chest. From the results, we found out that the segmentation error was around 2.17% on the average comparing with the expert's opinion. However, sometimes the system extracted smaller heart chambers. This might be because of the opening operator in the post processing stage.