This paper is leveraging the exceptional modeling capabilities of convolutional neural networks in an effort to build an close-to-close method where they immediately predict the ultrasound probe movement from the pictures themselves, and displays that this novel solution considerably outperforms the standard technique and it has direct scientific a