Analysis of Amniotic fluid in fetal assessment using Biophysical Profile Images

The assessment of prenatal Amniotic Fluid Volume (AFV) is critical because it provides essential information on fetal development and well-being, including perinatal prognosis. AFV measurement, on the other hand, is time-consuming and patient-specific. Furthermore, it depends on the sonographer, with measurement accuracy ranging substantially depending on the sonographer's experience. As a result, developing accurate, resilient, and adaptable methodologies to assess AFV is essential. Automation is expected to diminish user-based variability and the sonographer burden. Nevertheless, automation of AFV measurement is problematic because of several confounding variables, making precise recognition of AF pockets challenging. Furthermore, the forms and sizes of AF pockets are unknown, and ultrasound images frequently reveal them caused by missing structural boundaries. The suggested approach is to construct an Accurate and Efficient Deep Convolutional Neural Network (AEDCN-Net) predicated on an intensive preprocessing phase and a resourcefulness model structure to address the identified challenges. The AEDCN-Net uses asymmetric convolution-based residual skip interconnections in the processing path to learn feature representations with a broader receptive field while reducing the number of training data and floating-point operations (FLOPs). automatically generated when manuscripts are processed after acceptance.