A cone-beam X-ray computed tomography data collection designed for machine learning

• DOI: 10.1038/s41597-019-0235-y
• Publication Date: 2019-10-23T04:26:13Z
• Authors: Henri Der Sarkissian; Felix Lucka; Maureen van Eijna...; Giulia Colacicco; Sophia Bethany Coban; Kees Joost Batenburg
• Keywords: Ground truth; Data set

Unlike previous works, this open data collection consists of X-ray cone-beam (CB) computed tomography (CT) datasets specifically designed for machine learning applications and high cone-angle artefact reduction. Forty-two walnuts were scanned with a laboratory set-up to provide not only from single object but class objects natural variability. For each walnut, CB projections on three different source orbits acquired cone angles as well being able compute artefact-free, high-quality ground truth images the combined that can be used supervised learning. We complete image reconstruction pipeline: raw projection data, description scanning geometry, pre-processing scripts using software, reconstructed volumes. Due this, dataset reduction also algorithm development evaluation other tasks, such limited or sparse-angle (low-dose) scanning, super resolution, segmentation.