Computational method for optical diffraction tomography using a series of point sources

An algorithm based on the Fourrier transformation and deconvolution of integrated images for optical diffraction tomography is described. There are some methods to retrieve the three-dimensional object distribution such as an imaging with longitudal scanning, computed tomography based on Radon transformation, digital holographic method and so on. A series of variable angle of plane wave illumination is typically used for diffraction tomography. In this paper, diffraction tomography with a one-dimensional series of images illuminated by arbitrary profile of the light sources. The estimated three-dimensional object distribution is accumulated so that the accuracy becomes better. Since we assume to use arbitrary illumination profiles, the accumulated distribution has a distortion. However, is possible to compensate the object distribution by employing the calculated point spread function obtained through the same procedure. As examples, the conventional plane-wave illumination and the point source illumination cases are introduced in association with the effect of the accumulation and deconvolution with the point spread function to the accuracy of the calculated object distribution.