Effective implementation of ring artifacts removal filters for synchrotron radiation microtomographic images

The presence of ring artifacts in X-ray computed microtomography affects the qualitative and quantitative analyses of the reconstructed images. Although digital image processing approaches to ring artifacts removal via direct filtering of the reconstructed images exist, in synchrotron radiation microtomography this issue is usually faced during the actual reconstruction process by de-striping the sinogram image. In this work different de-striping algorithms are discussed and preliminary compared using a suitably created test image as well as actual imaged data. Because of the increasing need for fast reconstruction workflows, details of the implementation and the related computational aspects are also considered. Moreover, the hardware and software solution developed and deployed for the SYRMEP (SYnchrotron Radiation for MEdical Physics) microtomographic beamline of the Italian synchrotron radiation facility (Elettra - Sincrotrone Trieste S.C.p.A) is presented. This solution is based on a highperformance computing (HPC) cluster with the Oracle Grid Engine distributed resource management (DRM) system and Python code that takes advantage of the NumPy and SciPy libraries. This work makes available high resolution test images as well as source code at: ulisse.elettra.trieste.it/uos/pore3d/ringrem.