Model-based reconstruction for photoacoustic tomography with finite aperture detectors

Abstract

So far most rigorous reconstruction algorithms for photoacoustic tomography (PAT), for example, the modified back-projection algorithm, have been developed based on ideal point detectors. Nonetheless, because of the issue of poor signal-to-noise ratio from point-like detectors, finite-size flat transducers are commonly used in PAT, thus introducing the finite aperture effect - the tangential resolution in PAT deteriorates as the imaging point moves away from the circular scanning center. Such finite aperture effect and degradation in tangential resolution result from the spatial impulse response of the finite-size flat transducer. In this study, we adopt a model-based reconstruction method to improve the degraded tangential resolution for a photoacoustic tomography system with finite-size flat transducers. This method is based on a linear, discrete model of the photoacoustic detection system developed using matrix formalism. Using this model, a spatiotemporal optimal filter designed in minimum mean square error sense is used to deconvolve the spatial impulse responses associated with a finite-size flat transducer at every imaging point; thus retrospective restoration of the tangential resolution can be achieved. The performance of the model-based reconstruction method is verified using simulation data. It is demonstrated that the proposed method effectively improves the degraded tangential resolution for PAT with finite-size flat transducers while retaining the radial resolution. In addition, by taking electrical impulse response of the transducer into account, the proposed method can potentially improve the radial resolution, too.