In silico evaluation of the effect of sensor directivity on photoacoustic tomography imaging

Abstract

The effect of sensor directivity on photoacoustic tomography (PAT) image reconstruction is examined. An analytical expression for the resultant PA signal, when emitted by an ensemble of spherical sources and detected by a planar sensor with finite aperture, is derived. This framework was employed to calculate the PA signal produced by a vasculature phantom containing many randomly distributed spherical sources mimicking the red blood cells (RBCs). A Monte Carlo algorithm was employed to generate the random locations of RBCs inside the vasculature. The calculated signals for many detector locations were utilized for image reconstruction via the Tikhonov regularization method. The corresponding system matrix was constructed in three ways by considering (i) sensors as point detectors (PDs), (ii) finite detectors but without directivity (FD WOD), and (iii) finite detectors with directivity (FD WD). The novelty of this work is to incorporate the effect of sensor directivity during the signal simulation as well as system matrix formation. The FD WD algorithm provided accurate image reconstruction while PD and FD WOD failed to recover the vascular structure. Pearson’s correlation coefficient for the image formed using the FD WD technique was computed to be ≈ 4 times higher than that of the FD WOD method. The simulation results confirm that PAT imaging becomes independent of transducer directivity with the FD WD approach.