Reconstruction of Plane-Wave Imaging Using Dual-Domain Hybrid Neural Network

The image quality of plane wave (PW) ultrasound imaging is usually enhanced through Coherent Plane-Wave Compounding (CPWC), which utilizes multiple angles of PW for imaging. The image quality is strongly correlated with the number of PW angles, thus necessitating a trade-off between imaging time and image quality. Both radio frequency (RF) data and single-angle image reconstructed by Delay-and-Sum (DAS) contain rich information. Therefore, we designed a hybrid CNN-Transformer model that can leverage the advantages of the two types of data to improve the quality of single-angle PW images. Specifically, we designed the perception module to utilize the beamformed images as guidance to extract information from the RF signals. Compared to DAS, it achieved an improvement of 13.26 dB in CR. The experimental results demonstrated the performance of our proposed model.