Distance driven backprojection image reconstruction in digital tomosysthesis

In this paper, distance driven (DD) backprojection image reconstruction was investigated for digital tomosysthesis. Digital tomosysthesis is an imaging technique to produce three dimensional information of the object with low radiation dosage. This paper is a new study of DD backprojection for image reconstruction in digital tomosysthesis. Since DD considers that the image pixel and detector cell have width, the convolution operation is used to calculate DD coefficients and the approximation characteristics of some other methods such as ray driven method (RD) can be avoided. For three dimensional cone beam reconstruction, both projected voxel boundaries and detector cells boundaries are mapped to a common plane to estimate the overlap area. The overlap area is estimated by multiplying the overlap length along the (x or y) direction and along z direction. This paper represents a computer simulation result of DD with Maximum Likelihood Expectation Maximization (MLEM) of tomosysthesis reconstruction algorithms. The sequence of projection images were acquired from 25 projections with a total view angle of 48 degrees. DD with MLEM reconstruction results were demonstrated. The computation time in DD with MLEM to provide the reconstruction results is short since the main loop of DD is over x and y intercepts not over the image pixels or detectors cells compared with RD. In clinical applications, both the accuracy and speed of implementation condition are necessary requirements. DD backprojection can mostly satisfy the required conditions. In this paper, line profile along x direction was used to evaluate DD and RD methods.