activities

Research projects PSPC DIANA RHU ICELAND Student supervision postdocs phds interns Scientific animation The African Network for Artificial Intelligence in Biomedical Imaging Colloque Français d’Intelligence Artificielle en Imagerie Biomédicale

Colloque Français d'Intelligence Artificielle en Imagerie Biomédicale

Details I am part of the organizing committee for the 2024 conference. Website access The #IABM24 Conference is the French conference on artificial intelligence in biomedical imaging. It is organized by the three French 3IA institutes: MIAI Grenoble Alpes, 3IA Côte d’Azur, and PR[AI]RIE Institute. The IABM Conference aims to bring together French stakeholders in the field of artificial intelligence in biomedical imaging, whether from the academic, hospital, or industrial sectors.

Gudrun Olafsdottir

Prostate Segmentation in MR Images - A deep-learning approach for direct mesh reconstruction Master 2 BME Paris PSPC DIANA Description Traditionally, prostate volume segmentation in medical imaging has relied on voxel-based methods, which can be computationally expensive and less accurate in certain cases. The goal of this internship was to propose a novel method based deep learning segmentation on 3D meshes. Mesh-based segmentation offers several advantages, including efficient processing, representation of complex geometries, and potential for integration with clinical expertise.

interns

2022 Thi Huong Ly Nguyen Houda Guerwane 2023 Gudrun Olafsdottir Pascal Zhang 2024 Marie Edet Paul Gornet Rémi Gendron

Pascal Zhang

Automatic Detection of Fiducials in Prostate Images using Deep Learning Master 2 Phelma - Biomed Welcome Grant UGA

phds

Tamara Dupuy Louis Lenfant Chloé Soormally

postdocs

Thao Ho Thi

PSPC DIANA

Brief Aims at developing new systems to assist in the management of prostate cancer. Research objectives In this project, we propose to redesign several stages of the current ultrasound platform offered by Koelis. The goal is to increase the speed and reliability of existing processes by introducing deep learning models. The studied applications include gland segmentation in MRI and TRUS imaging (see Louis Lenfant), deformable multimodal registration of MRI and TRUS (see Thao Ho Thi), and navigation assistance (targeting) during biopsy (see Tamara Dupuy, Thi Huong Ly Nguyen, Chloé Soormally).

RHU ICELAND

Brief Intra Cardiac Echography integration in Landing an Annuloplasty Device for transcatheter mitral valve repair. (Website access) Funding ANR France Partners DigiSurge Hospices Civils de Lyon UCBL Lamcos & Mateis teams (INSA Lyon) Central Lyon (LTDS) Sorbonne University (ISIR) CHU St Etienne TIMC

Tamara Dupuy

Navigation en temps-réel de biopsie prostatique assistée par intelligence artificielle MIAI As part of her thesis, Tamara focused on navigation during the complex procedure of prostate biopsy. This procedure is currently performed without assistance (mental mapping by the clinician), and one way of improving the quality of this clinical gesture would be to assist the urologist during navigation towards a target to be sampled, which is moreover only visible on MRI.

Thao Ho Thi

Multimodal 3D deformable MRI / TRUS registration using deep learning #PSPC DIANA

The African Network for Artificial Intelligence in Biomedical Imaging

Details I am co-mentoring two African students for the #AFRICAI project (Website access) organized for the MICCAI 2024 conference.

Thi Huong Ly Nguyen

Segmentation-based registration for real-time prostate biopsy navigation Master 2 Phelma - Biomed PSPC DIANA Description The goal of this internship was to work on a parallel branch to the work of Tamara Dupuy . The idea was to propose an extension to the 2D/3D registration model, by considering the possibility of introducing an additional input corresponding to the prostate segmentation into the initial model. Several architectures were tested and evaluated in the context of this internship.