On discrete X-ray transform
Roman Novikov, Basant Lal Sharma
TL;DR
This work analyzes a discrete version of the X-ray transform on $\mathbb{Z}^d$ with the goal of non-overdetermined reconstruction from limited data. It distinguishes irrational and rational ray directions, establishing direct pointwise inversions for irrational rays via $f(x)=\mathcal{P}f(\gamma_{x,\theta})$ and a slice-by-slice, Cormack-type reconstruction for rational rays on structured sets $\mathpzc{T}^*$ (and extensions $\mathpzc{T}^*_{\alpha,\beta}$) under $\mathrm{supp}\,f \subset B_r$, with results that extend to weighted versions $\mathcal{P}_W$. The paper also clarifies the relationship between discrete and continuous X-ray transforms, showing exact correspondences for lattice delta inputs and deriving regular-case relations that connect $\mathcal{P}^{\text{con}}$ and $\mathcal{P}^{\text{dis}}$ via residuals like $\delta f$; these bridges provide a framework to transfer insights between continuous tomography and its discrete sampling. Overall, the results deliver non-overdetermined, staircase-like reconstruction procedures for both irrational and rational ray families, include weighted generalizations, and illuminate the discrete-to-continuous transition with potential computational benefits in tomographic sensing.
Abstract
We consider a discrete version of X-ray transform going back, in particular, to Strichartz (1982). We suggest non-overdetermined reconstruction for this discrete transform. Extensions to weighted (attenuated) analogues are given. Connections to the continuous case are presented.
