Photoacoustic image reconstruction from a frequency-invariant source localization perspective

Abstract

Photoacoustic imaging provides high spatial resolution images of biological tissues and is useful for molecular imaging. The exact reconstruction algorithms for photoacoustic imaging are either slow or assume a continuous sensor with infinite bandwidth. We propose a novel reconstruction method which expands the source distribution function in the Fourier-Bessel domain. The source distribution can be reconstructed from frequency samples corresponding to the Bessel zeros. Sparsity of the source distribution in the Fourier-Bessel domain makes reconstruction faster. Further, this method was extended to the discrete aperture and a condition was derived to avoid spatial aliasing. The proposed method was verified using numerical simulations.