Ultrasound Imaging of Bone Cortex Using Block-Based Reconstruction

Without considering the significant acoustic impedance contrast between bone and soft tissue, traditional imaging methods with unique sound velocity assumption are challenging to generate accurate ultrasound image of bone cortex, such as time-domain synthetic aperture (TDSA), and phase shift migration (PSM). Furthermore, TDSA restores the image point-by-point, and PSM reconstructs the target layer-by-layer, leading to the relatively high computation cost. To overcome these limitations, this work proposed a block-based fast ultrasound imaging method using velocity model estimation (BR-FUI-VE). After the estimation of sound velocity model via travel-time inversion, the imaging process involved block-by-block reconstruction, merely extrapolating the received wavefield to the top of the target block. This facilitated the reconstruction of entire tissue block (e.g., soft tissue or bone tissue) in a single step through 2D fast Fourier transform (2D-FFT). The effectiveness of the method was demonstrated by one simulated specimen, two cortical phantoms, and two ex-vivo goat tibias. The estimated sound velocity model showed mean relative errors below 14%, and the mean errors of the cortical thickness were less than 0.32 mm. The results of ex-vivo experiments were in good agreement with the reference models measured by micro computed tomography (μCT). Moreover, BR-FUI exhibited significantly reduced time complexity compared to TDSA and PSM, allowing a 5-layer image (i.e., goat tibia) to be reconstructed in just 0.3 seconds. Unlike the conventional synthetic aperture (SA), the proposed BR-FUI-VE method was proven to be an effective modality for accurate and efficient cortical bone imaging.