Experimental benefits display that equally anatomical and protocol variances are enabling factors for DNN-dependent US reconstruction; and learning tips on how to discriminate distinct subjects and predefined different types of scanning paths both substantially boost frame prediction accuracy, volume reconstruction overlap, accrued tracking mistake