Table 3.

State-of-the-art literature on the DNN for localization of anatomical structure.

Reference	Adopted DNN	Dataset used	Research findings
Baumgartner et al. [79]	CNN	US video frame data with 1003 midpregnancy scans	Precision, 69% and recall, 80%,
Ghesu et al. [15]	Adaptive DNN with marginal space learning	3D US images collected from 869 patients with 2891 volumes	Accuracy, 45.2%
Chen et al. [77]	CNN with transfer learning	11,942 foetal abdominal US images collected from pregnant women	Accuracy, 89.6%
Chen et al. [26]	Extended (Chen et al., 2015b) with LSTM to extract spatial temporal features	11,942 foetal abdominal US images collected from pregnant women	Accuracy, 90.8%
Baumgartner et al. [81]	CNN based on a VGG-16 model	2694 2D US images taken from gestational women	Accuracy, 77.8%