ImFusion Suite: GPU-based Cone-beam computed tomographic (CT) reconstruction

This video presents our revamped GPU-based Cone-beam computed tomographic reconstruction. Our software includes:
1) Projection preprocessing such as flat-field correction, log conversion, and dead-pixel masking.
2) Geometric calibration for Cone-beam systems as well as direct support for several vendors.
3) GPU based tomographic reconstruction which supports both FDK, and iterative algorithms.
4) Artifact reduction methods for metal, beam-hardening, and rings.
5) Support for post-processing such as HU calibration.

The dataset was taken from the VerSe2020 dataset:
1. Löffler M, Sekuboyina A, Jakob A, Grau AL, Scharr A, Husseini ME, Herbell M, Zimmer C, Baum T, Kirschke JS. A Vertebral Segmentation Dataset with Fracture Grading. Radiology: Artificial Intelligence, 2020 https://doi.org/10.1148/ryai.2020190138.
2. Liebl H, Schinz D, Sekuboyina A, ..., Kirschke JS. A computed tomography vertebral segmentation dataset with anatomical variations and multi-vendor scanner data Sci Data. 2021 Oct 28;8(1):284. doi: 10.1038/s41597-021-01060-0.
3. Sekuboyina A, Bayat AH, Husseini ME, Löffler M, Menze BM, ..., Kirschke JS. VerSe: A Vertebrae labelling and segmentation benchmark for multi-detector CT images. Med Image Anal. 2021 Oct;73:102166. doi: 10.1016/j.media.2021.102166. Epub 2021 Jul 22. preliminary access at https://arxiv.org/abs/2001.09193