"Fourier-Domain Inversion for the Modulo Radon Transform"Beckmann, MatthiasInspired by the multiple-exposure approach in computational photography, in recent years, practitioners have explored the idea of high dynamic range (HDR) X-ray imaging and tomography to deal with high intensity projections that exceed the dynamic range of the detector. While yielding promising results, these approaches inherit the limitations of the multiple-exposure fusion strategy. To overcome these disadvantages, in our recent line of work the modulo Radon transform has been proposed, which is based on a co-design of hardware and algorithms. In the hardware step, Radon transform projections are folded using a modulo non-linearity to avoid sensor saturation. Thereon, recovery is performed by algorithmically inverting the folding, thus enabling a single-shot, HDR approach to tomography. In this talk we propose a novel Fourier-domain recovery algorithm that is based on the Orthogonal Matching Pursuit (OMP) algorithm as well as the non-equispaced Fast Fourier Transform (NFFT). The advantages of our approach include recovery at lower sampling rates while being agnostic to the modulo threshold, lower computational complexity and empirical robustness to noise. In addition to numerical simulations we employ prototype modulo ADC based hardware experiments to validate our claims, where we report image recovery based on hardware measurements up to 10 times larger than the sensor's dynamic range. This talk is based on joint work with Ayush Bhandari (Imperial College London) and Meira Iske (University of Bremen). |
« back