Synergistic Multi-Energy CT Reconstruction with a Deep Penalty “Connecting the Energies”

Zhihan Wang (1) , Alexandre Bousse (1) , Noel Jeffrey Pinton (1) , Jacques Froment (2) , Franck Vermet (2) , Béatrice Vedel (2) , Jean-Pierre Tasu (3) , Dimitris Visvikis (1)

1 LaTIM - Laboratoire de Traitement de l'Information Medicale
2 LMBA - Laboratoire de Mathématiques de Bretagne Atlantique
3 POITIERS - Radio - Service de Radiologie

We propose a novel penalty term for multi-channel synergistic image reconstruction with an application to multi-energy computed tomography (CT). The penalty utilizes trained convolutional neural networks (CNNs) to connect the energies to a latent image. We show on simulated data that our method has the potential to outperform reconstruction with a joint total variation (JTV) penalty.

Domaines: Imagerie médicale; Traitement des images [eess.IV]

Zhihan Wang, Alexandre Bousse, Noel Jeffrey Pinton, Jacques Froment, Franck Vermet, et al.. Synergistic Multi-Energy CT Reconstruction with a Deep Penalty “Connecting the Energies”. 2022 IEEE Nuclear Science Symposium (NSS), Medical Imaging Conference (MIC) and Room Temperature Semiconductor Detector (RTSD) Conference, IEEE, Nov 2022, Milan, Italy. ⟨hal-03955092⟩