A robust modified delay-and-sum algorithm for photoacoustic tomography imaging with apodized transducers

Role of sensor sensitivity in photoacoustic tomography (PAT) imaging is discussed. In this study, sensitivity profile (apodization) of finite size sensors was considered as axisymmetric and modelled by using a Gaussian function. The full width at half maximum (FWHM) of the Gaussian function was varied in order to investigate its effect on PAT image reconstruction. The images were reconstructed using conventional delay-and-sum (CDAS) and modified delay-and-sum (MDAS) algorithms. In case of the CDAS, a Gaussian function was used to weight the PA signals detected by different parts of the sensor and the resultant signal was computed by summing those signals. However, in case of the MDAS, the Gaussian weight was applied in both directions (signal acquisition and redistribution of the pressure values at different point locations on the aperture of the finite sensor). The performance of these algorithms was investigated with respect to ideal point detectors by conducting numerical experiments in the k-Wave toolbox. The results for the CDAS and the MDAS algorithms are found to be very close to that of ideal point detectors when FWHM is small. The MDAS technique appears to be much superior to the CDAS approach when FWHM is large. The MDAS method can be employed in practice for apodized transducers as well if the Gaussian weight is applied in both directions (signal acquisition and redistribution).