Four-dimensional wide-field ultrasound reconstruction system with sparse respiratory signal matching


Abstract—Four-dimensional (4D) ultrasound reconstruction can greatly extend the spatial and temporal range of two-dimensional (2D) ultrasound in clinical practice. However, uneven breaths may yield a considerable motion artifact in the reconstructed time sequences of volume ultrasound. In this paper, a systemwith sparse respiratory signal matching is proposed to realize accurate 4D ultrasound reconstruction by effectively estimating uneven breaths. A slippery platform is built to automatically control the ultrasonic probe for acquiring 2D ultrasound slice sequences in different abdominal areas. Based on these acquired sequences, the signals of the respiratory phases are first extracted. Second, the sparse matched signal pairs are determined through the signal distribution and the corresponding image connectivity. After matching is conducted, signals in different positions are aligned in a common space wherein the outliers of the signals caused by the uneven breaths are excluded. Third, images with the same signal value in different positions are collected as the 3D image at the phase corresponding to the signal, and the 3D images at different phases are employed to reconstruct the final 4D image. The accuracies of the reconstructed 4D image are evaluated and compared with those of five existing methods. Experimental results demonstrate that the spatial and temporal continuities of the 4D ultrasound image reconstructed by our method outperform those obtained by other methods.