An analytical method for 3D tomographic reconstruction from a small number of projections using a simple interpolation scheme

A new algorithm is presented for three-dimensional tomographic reconstruction from a reduced number of projections. It is based on the completion of data by interpolation of the measured data in the frequency domain. A filtered backprojection procedure, suited to conic geometry, performs the tomographic reconstruction from the augmented data set. A second reconstruction scheme is derived for high-contrast sparse objects. It improves the resolution by zeroing the positives values of the filtered projections before the backprojection. These methods have been tested on simulated data and data obtained on a 3-D X-ray acquisition system. They produce better quality 3-D images than analytical reconstruction working alone