In Vivo Three-Dimensional Raster Scan Optoacoustic Mesoscopy Using Frequency Domain Inversion

Abstract

Optoacoustic signals are typically reconstructed into images using inversion algorithms applied in the time-domain. However, time-domain reconstructions can be computationally intensive and therefore slow when large amounts of raw data are collected from an optoacoustic scan. Here we consider a fast weighted -(FWOK) algorithm operating in the frequency domain to accelerate the inversion in raster-scan optoacoustic mesoscopy (RSOM), while seamlessly incorporating impulse response correction with minimum computational burden. We investigate the FWOK performance with RSOM measurements from phantoms and mice in vivo and obtained 360-fold speed improvement over inversions based on the back-projection algorithm in the time-domain. This previously unexplored inversion of in vivo optoacoustic data with impulse response correction in frequency domain reconstructions points to a promising strategy of accelerating optoacoustic imaging computations, toward video-rate tomography.

Frequency wavelength multiplexed optoacoustic tomography

“…Therefore, unprocessed images obtained directly from the system are heavily blurred, and require further processing to obtain high-contrast and high-resolution images. To do so, a back-projection algorithm is implemented in the Fourier domain 30 , 31 to recover an (acoustically) diffraction-limited image. The resulting image is then corrected by the system impulse response 32 and further processed with a vesselness filter for display purposes 33 .…”

Enabling the autofocus approach for parameter optimization in planar measurement geometry clinical optoacoustic imaging

“…Common implementations of time-domain backprojection reconstruction [15], which involves a complexity of O(n 5 ), generally take around 2 min to perform one reconstruction and therefore around 1.5 h to calculate the AF algorithm. Even frequency-domain backprojection [15], which has a lower computational complexity of O(n 3 log 2 [n]), requires approximately 15 s per reconstruction [16] and therefore around 15 min to calculate the AF algorithm. Reconstruction times are even longer for other image formation approaches, such as model-based reconstruction [17][18][19].…”

“…As a result, the AF algorithm can be correctly applied to single B-scans by calculating a sharpness metric for the 2D reconstructions of the virtual sources (see Section 2 and Figure 1A). This process substantially reduces the computational complexity of image formation from O(n 5 ) to O(n 3 ) in the case of time-domain reconstructions, and from O(n 3 log 2 [n]) to O(n 2 log 2 [n]) in the case of frequency-domain reconstructions [15] (also known as w-k algorithms [16]). This reduction in complexity translates to faster computation of the FAF algorithm.…”