Ander Biguri

and 21 more

X-ray computed tomography (CT) is an important noninvasive medical imaging modality for discovering the structural details of internal organs. The image reconstruction task in CT is an inverse problem that seeks to recover the internal structure of an object by measuring the absorption profile of X-ray beams (sinogram) measured using a detector. The classical variational approach for CT image reconstruction is to minimize an energy function using an appropriate iterative optimization algorithm. Motivated by the success of deep learning, researchers in recent years have begun to leverage training data and enhanced computing capabilities to produce high-fidelity reconstructed images. Nonetheless, much of the academic research in deep learning algorithms for CT has focused primarily on the two-dimensional setting (with simplified forward operators and noise model) for demonstrating proofs-of-concept, and a comprehensive benchmarking of various classical and data-driven CT reconstruction approaches has not been undertaken. The key objective of our CT reconstruction grand challenge is to promote methodological advancements for both classical and deep learning-based approaches for clinical CT with a reasonably accurately simulated 3D CT forward operator and noise model. We have utilized the publicly available LIDC-IDRI dataset and simulated sinograms and FDK images corresponding to two dose levels (clinicaland low-dose, constituting two tracks of the challenge) starting from the normal-dose images as the ground truth. In this paper, we summarize the motivation, context, and results of our challenge, and highlight the future research directions in deep learning for clinical CT.