Supervision : Andrea PINNA

Co-supervision : BLOCH Isabelle (LIP6) - SZEWCZYK Jérôme (ISIR)

Image analysis with artificial intelligence methods for digestive endoscopy, ERCP

Endoscopic retrograde cholangiopancreatography or ERCP is a minimally invasive procedure aiming at removing stones or tumors from biliary duct. This operation takes place under 2D radiographic control to verify the location of surgical instruments such as catheter, guidewire etc. The 3D anatomy of the biliary tree and the great interpatient variability make this act, however minimally invasive, very complex to perform.
The radio opacity of the instruments is not sufficient for the localization of the latter. This is why the use of contrast agent is necessary. However, this exposes the patient to complications which can prove to be serious.
The aim of this thesis is to propose reproducible and explainable artificial intelligence methods for the localization of surgical instruments in the biliary tree.
For this, we will rely in particular on the results obtained previously as well as the use of mathematical morphology for the modeling of the topology of the instruments.

2022 Publications

• 2022
  • O. Thiery, G. Martin, I. Bloch, M. Dres, U. Saleem, A. Pinna : “Semi-Automatic A-Line Detection and Confidence Scoring in Lung Ultrasound”, BioCAS 2022 - Biomedical Circuits and Systems Conference, Taipei, Taiwan, Province of China (2022)
  • G. Martin, S. El‑Madafri, A. Becq, J. Szewczyk, I. Bloch : “Instruments Segmentation in X-ray Fluoroscopic Images for Endoscopic Retrograde Cholangio Pancreatography”, Studies in Health Technology and Informatics, vol. 294, Studies in Health Technology and Informatics, pp. 133-134, (IOS Press) (2022)