Placental maturity grading via hybrid descriptors based on fisher vector

Abstract

Placental maturity grading (PMG) is quite essential to assess fetal growth and maternal health. To this date, PMG has mostly relied on the clinician's subjective judgment, which is time-consuming and subjective. Traditional machine learning-based methods capitalize on handcrafted features, but such features may be essentially insufficient for PMG. In order to tackle it, we propose an automatic method to stage placental maturity via deep hybrid descriptors based on B-mode ultrasound (BUS) and color Doppler energy (CDE) images. Specifically, convolutional descriptors extracted from a deep convolutional neural network (CNN) and handcraft features are combined to form hybrid descriptors to boost the performance of the proposed method. First, different models with various feature layers are combined to obtain hybrid descriptors from images. Meanwhile, the transfer learning strategy is also utilized to enhance the grading performance based on the deep representation features. Then, extracted descriptors are encoded by Fisher vector (FV). Finally, we use support vector machine (SVM) as the classifier to grade placental maturity. The experimental results demonstrate that our proposed method has achieved remarkable staging performance.