Accurate Detection of Septal Defects With Fetal Ultrasonography Images Using Deep Learning-Based Multiclass Instance Segmentation

Siti Nurmaini,Annisa Darmawahyuni,Ade Iriani Sapitri,Bambang Tutuko,Adithia Jovandy,Firdaus Firdaus,Muhammad Naufal Rachmatullah,Rossi Passarella

doi:10.1109/access.2020.3034367

Siti Nurmaini, Annisa Darmawahyuni + Show 6 more

Open Access

https://doi.org/10.1109/access.2020.3034367

Copy DOI

Abstract

Accurate screening for septal defects is important for supporting radiologists' interpretative work. Some previous studies have proposed semantic segmentation and object detection approaches to carry out fetal heart detection; unfortunately, the models could not segment different objects of the same class. The semantic segmentation method segregates regions that only contain objects from the same class. In contrast, the fetal heart may contain multiple objects, such as the atria, ventricles, valves, and aorta. Besides, blurry boundaries (shadows) or a lack of consistency in the acquisition ultrasonography can cause wide variations. This study utilizes Mask-RCNN (MRCNN) to handle fetal ultrasonography images and employ it to detect and segment defects in heart walls with multiple objects. To our knowledge, this is the first study involving a medical application for septal defect detection using instance segmentation. The use of MRCNN architecture with ResNet50 as a backbone and a 0.0001 learning rate allows for two times faster training of the model on fetal heart images compared to other object detection methods, such as Faster-RCNN (FRCNN). We demonstrate a strong correlation between the predicted septal defects and ground truth as a mean average precision (mAP). As shown in the results, the proposed MRCNN model achieves good performance in multiclass detection of the heart chamber, with 97.59% for the right atrium, 99.67% for the left atrium, 86.17% for the left ventricle, 98.83% for the right ventricle, and 99.97% for the aorta. We also report competitive results for the defect detection of holes in the atria and ventricles via semantic and instance segmentation. The results show that the mAP for MRCNN is about 99.48% and 82% for FRCNN. We suggest that evaluation and prediction with our proposed model provide reliable detection of septal defects, including defects in the atria, ventricles, or both. These results suggest that the model used has a high potential to help cardiologists complete the initial screening for fetal congenital heart disease.

Highlights

Congenital heart diseases (CHDs) are the most common malformations and occur in 0.8% of the general population [1]
To ensure that the proposed deep learning-based MRCNN architecture can work properly, four stages are proposed: (i) preconfigured bounding boxes are with various image shapes, and resolutions are established; (ii) the highest boundary boxes are defined to generate regional proposals; (iii) composite region proposals are pruned using non-maximum suppression and used to determine the presence or absence of a hole in a septum; and (iv) segmentation masks are produced for cases in which septal defects are positive, i.e., atrial septal defects (ASDs), ventricular septal defects (VSDs), and atrioventricular septal defects (AVSDs)
The markers are placed in the wall chamber, aorta, and hole for ASDs, VSDs, AVSDs, and normal conditions

Summary

INTRODUCTION

Congenital heart diseases (CHDs) are the most common malformations and occur in 0.8% of the general population [1]. The automated segmentation of the fetal standard scan plane for detecting a hole in the septum has never been performed in this challenging scenario This is the first study in a medical application for septal defect detection using MRCNN. To ensure that the proposed deep learning-based MRCNN architecture can work properly, four stages are proposed: (i) preconfigured bounding boxes are with various image shapes, and resolutions are established; (ii) the highest boundary boxes are defined to generate regional proposals; (iii) composite region proposals are pruned using non-maximum suppression and used to determine the presence or absence of a hole in a septum; and (iv) segmentation masks are produced for cases in which septal defects are positive, i.e., ASDs, VSDs, and AVSDs. In all stages, as seen, hole detection is an essential component of this study for making septal defect decisions. Where APi is the AP in the ith class and N is the total number of classes being evaluated

RESULTS AND DISCUSSION

Method

CONCLUSION