Inverse Radon transform for optoacoustic imaging

Publisher Website
Google Scholar
Abstract
Mathematical model of image reconstruction for two-dimensional optoacoustic imaging system is described. It was assumed that receiving transducers are uniformly distributed along the perimeter of a 60 mm radius ring with 2.1 mm gaps between transducer centers and initial data were known with 0.1 mm increments. The algorithm of radial back projection with convolution was used for optoacoustic image reconstruction. The convolution was evaluated with modified Shepp-Logan (MSL) and rectangular (RECT) space spectrum windows. Linear interpolation was applied for calculation of the convolution at the intermediate space points. The following four criteria were employed for estimation of resulting image quality: noise level on entire tomogram, a jump transfer function, loss contrast function and the contrast-dimension reflation. Theoretical expressions for these parameters were derived and used for optimization of the proposed algorithm. Two-dimensional images of computer simulated spherical objects were reconstructed. It was shown that 0.1 mm spatial resolution could be obtained provided the signal-to-noise ratio equals approximately 3 at the tomogram. A very small (0.2 mm diameter) tumor and a small (2-mm diameter) tumor could be clearly revealed at the tomogram if their optical absorption contrast equals at least 2 and 0.1 respectively.

This publication has 0 references indexed in Scilit: