Deep Learning-based 3D Beamforming on a 2D Row Column Addressing (RCA) Array for 3D Super-resolution Ultrasound Localization Microscopy

We report 3D ULM imaging by using a 2D row-column addressing (RCA) array which achieves fast imaging volume rate. Furthermore, we propose a deep-learning (DL)-based adaptive beamforming method to improve the spatial resolution and contrast of the microbubble (MB) signal imaged by the RCA array. We evaluated the proposed technique on a wire phantom and MBs suspended in water. Moreover, we carried out an in vivo study on a mouse brain and demonstrated improved 3D ULM imaging based on the DL-beamformer. These results demonstrate that DL-based beamforming provides a viable solution for enhancing the RCA imaging quality for robust ULM.