An Overview of Non-invasive Detection of Fetal Abnormalities Using Deep Learning Techniques

R. Ramya, K. Srinivasan, V. Dharshini, T. Immanuel Johnson and N. Varsha

Deep learning technique is great interest in medical imaging. Motive of this proposed work is that advancing the state of the art in machine learning in medical image analysis. Abnormalities present in amniotic fluid volume causes of perinatal mortality and morbidity. Parameters to be access the fetal growth and development.An ultrasound (US) image of the abdomen is segmented into four equal coordinates and the sum of these quadrants resulting in Amniotic Fluid Index (AFI). Growth restriction is observed in 30 million infants every year and there is an association between oligohydramnios (less amniotic fluid) and both Intrauterine Growth Restriction and increased perinatal mortality. Performance of Adaptive Neuro-Fuzzy Inference classifier is applied for classification approach. The work is focused to detect Amniotic Fluid Volume automatically based on, deep learning for semantic segmentation with Fully Connected Convolution Neural Network for the implementation. The role of this approach is to detect malformation during the second and third trimester of pregnancy. This approach is to reduce the diagnosis time and the risk factors in previous stages of pregnancy using Machine learning techniques.

Volume 12 | Issue 8

Pages: 01-13

DOI: 10.5373/JARDCS/V12I8/20202442