A uniqueness result for the inverse problem of identifying boundaries from weighted Radon transform (Q6085053): Difference between revisions

The weighted Radon transform in \(\mathbb{R}^n\) is given by \[ ({\mathcal U} f)(x,\omega) = \int_{(y-x) \cdot \omega = 0} V(x,y) f(y) dy, \] where \(x, y, \omega \in \mathbb{R}^n\), \(\omega\) a unit vector; \(V(x,y)\) is a weight function, with \(V(x,x) \ne 0\). We assume that the support of \(f\) is a region \(G\) consisting of disjoint subregions \(G_i\) with continuous boundaries, with some restrictions on the geometry, and that \(f\) is Hölder continuous on each subregion, with jumps between subregions. It is shown that if \(n = 2m+1\) is odd, then the discontinuities of \(f\) (that is, the subregion boundaries) can be determined from \(\mathcal{U} f\). The main result basically states that the discontinuities of \(f\) are the same as the discontinuities of the \(m\)th power of the Laplacian applied to the (unweighted) backprojection of \(\mathcal{U} f\).