Coherence-weighted beamforming and automated vessel segmentation for improving photoacoustic imaging of embryonic vasculature using annular arrays

This paper presents an adaptive synthetic-focusing scheme for improving the quality of high-frequency ultrasound (HFU) and photoacoustic (PA) images acquired using a 40-MHz, 5-element, annular-array transducer. A 532-nm laser beam was coaxially delivered through an 800-μm hole in the central element of the array. The assembly was raster-scanned to acquire co-registered, 3D HFU and PA image data. Two synthetic-focusing schemes were compared: delay-and-sum (DAS) and DAS-weighted with an inter-channel coherence factor (DAS+CF). Bench-top experiments that employed an 80-μm hair and a tissue-mimicking phantom embedded with 530-μm anechoic spheres were performed to compare the two schemes applied to PA and HFU data, respectively. The DAS+CF scheme increased the SNR by approximately 10 dB for both modalities. The dimension of the hair in DAS-only PA image with a 20-dB dynamic range was approximately twice that in the corresponding DAS+CF image. The DAS-only HFU image of the anechoic phantoms exhibited a depth-dependent contrast ratio (CR) of less than 0.7 beyond the depth of 14 mm (CR = 1 represents a perfect contrast), whereas the CR remained greater than 0.7 in the DAS+CF image. DAS+CF approach also improved visualization of brain ventricles and vasculature in a 12-day old mouse embryo.