Background, Motivation and ObjectiveUltrafast imaging based on plane wave (PW) / diverging wave (DW) is an active area of research in ultrasound acquisitions because of its capacity of reaching high frame rate. Beamforming of received echoes in Fourier domain provides comparable image quality but lower computational complexity with respect to the conventional DAS approach [Zhang et al., UFFC, 2016]. However, the current state-of-the-art Fourier-based techniques only perform the reconstruction on radio frequency (RF) data. To process less data without loss of information, we adapted in this study the Fourier-based formalism of Lu [Lu et al., UFFC, 1997] to IQ data reconstruction for both PW and DW imaging.Statement of Contribution/MethodsThe original PW imaging method of Lu performs 2D FFT on the received RF signals and remaps the echoes spectrum to obtain the object spectrum, followed by a 2D IFFT to get the beamformed image. We demonstrated in this work that this method can be adapted for IQ data reconstruction by explicitly taking into account the demodulation frequency in thespectrum remapping relation. The proposed method can be easily extended to DW imaging from IQ data by using a space transformation introduced in one of our previous work [Zhang et al., UFFC, 2016]. However, the space transformation introduces time shifts that cause phase rotation and thus degrade compounding strategies. We thus integrated thosetime shifts into our formalism to compensate for the phase rotation during the reconstruction process.Results/DiscussionThe proposed approach is validated experimentally with an ultrasound phantom (CIRS model: 054GS) with PW / DW insonifications using a standard linear / phased array interfaced with a Verasonics system. The RF data was demodulated and downsampled 4 times to obtain the IQ data. Figure1 shows B-mode images reconstructed from RF data and IQ datausing 9 PW / 15 DW in transmission. This results show the feasibility of performing IQ data reconstruction in the Fourier domain. It allows reconstructing the same image quality as from RF data but with 4 times less data samples, thus the beamforming process could be increased 4 times faster.