This paper presents an ultrasound (US) volume reconstruction method only from US image sequences using deep learning.
The proposed method employs a convolutional neural network to estimate the position of a 2D US probe only from US images.
We introduce a network for estimating motion and two loss functions: the $L_1$ loss function and the consistency loss function.
Through a set of experiments using US image sequence datasets with ground-truth motion measured by a motion capture system, we demonstrate that the proposed method exhibits better performance on probe localization and volume reconstruction compared with the conventional method.