Preconditioned P-ULA for Joint Deconvolution-Segmentation of Ultrasound Images

Marie-Caroline Corbineau (1) , Denis Kouamé (2) , Emilie Chouzenoux (1) , Jean-Yves Tourneret (2, 3, 4) , Jean-Christophe Pesquet (1)

1 GALEN-POST - Organ Modeling through Extraction, Representation and Understanding of Medical Image Content
2 IRIT - Institut de recherche en informatique de Toulouse
3 ENSEEIHT - Ecole Nationale Supérieure d'Electrotechnique, d'Electronique, d'Informatique, d'Hydraulique et de Télécommunications
4 TéSA - Télécommunications Spatiales et Aéronautiques - Telecommunications for Space ant Aeronautics

Joint deconvolution and segmentation of ultrasound images is a challenging problem in medical imaging. By adopting a hierarchical Bayesian model, we propose an accelerated Markov chain Monte Carlo scheme where the tissue reflectivity function is sampled thanks to a recently introduced proximal unadjusted Langevin algorithm. This new approach is combined with a forward-backward step and a preconditioning strategy to accelerate the convergence, and with a method based on the majorization-minimization principle to solve the inner non-convex minimization problems. As demonstrated in numerical experiments conducted on both simulated and in vivo ultrasound images, the proposed method provides high-quality restoration and segmentation results and is up to six times faster than an existing Hamiltonian Monte Carlo method.

Mots clés: Ultrasound Imaging; Markov chain Monte Carlo MCMC methods; Proximity operator; Deconvolution; Segmentation

Domaines: Traitement des images [eess.IV]; Imagerie médicale; Optimisation et contrôle [math.OC]

Marie-Caroline Corbineau, Denis Kouamé, Emilie Chouzenoux, Jean-Yves Tourneret, Jean-Christophe Pesquet. Preconditioned P-ULA for Joint Deconvolution-Segmentation of Ultrasound Images. [Research Report] CVN, CentraleSupélec, Université Paris-Saclay, Gif-Sur-Yvette, France. 2019. ⟨hal-02073283v1⟩