Quantification de la qualité d'un geste chirurgical à partir de connaissances a priori

Rémi Wolf 1
1 TIMC-IMAG-GMCAO - Gestes Medico-chirurgicaux Assistés par Ordinateur
TIMC-IMAG - Techniques de l'Ingénierie Médicale et de la Complexité - Informatique, Mathématiques et Applications, Grenoble - UMR 5525
Abstract : The development of laparoscopic surgery has led to new challenges for surgeons, their visual and tactile perception of the operating field having been modified compared to their prior experience. Numerous devices have been designed around the surgical process in order to help the surgeon to perform the best possible intervention. These devices aim at enhancing the surgeon's perception of the operating context, optimally planning the surgical strategy, and assisting him/her during the intervention. The design of a system dedicated to the analysis of the surgical process, enabling the identi cation of risky situations and improving the quality of the surgery, is a major issue in the field of Computer Assisted Medical Interventions. The assessment of quality in surgery covers different aspects : the technical skills of the surgeon, as well as his/her theoretical knowledge and decision-making abilities. The objective of this thesis was to develop a technical skills assessment device for laparoscopic surgery, using a priori knowledge, suitable for the operating room's specific constraints without modifying the surgeon's environment. This assessment is based on the definition of predictive metrics for the quality of the surgery, derived from the instruments' trajectories during the procedure. The first step of this work consisted in the implementation of an instrument tracking method, based on endoscopic images without addition of any markers, in order to retrieve the 3D position of the instrument's tip inside the abdominal cavity during the surgical procedure. This tracking combines geometric models of the camera, the instrument and its orientation, with statistical models describing the evolutions of the latter. This method allows for simultaneous off-line tracking of multiple instruments within a training bench environment. The second step consisted in the extraction, from these trajectories, of predictive metrics for the assessment of the technical skills of the surgeon, using partial least squares regression and k-means classifiers. Several new metrics were identified, relating to the coordination of the surgeon's hands and the optimization of his/her workspace. This device is intended to be integrated in a more general system, in order to provide the surgeon with context-aware information regarding the surgical process, for example by merging data obtained from the trajectory with per-operative multi-modal image data.
Complete list of metadatas

Cited literature [180 references]  Display  Hide  Download

Contributor : Rémi Wolf <>
Submitted on : Monday, March 24, 2014 - 5:13:34 PM
Last modification on : Tuesday, September 24, 2019 - 4:22:03 PM
Long-term archiving on: Tuesday, June 24, 2014 - 12:45:41 PM


  • HAL Id : tel-00965163, version 1



Rémi Wolf. Quantification de la qualité d'un geste chirurgical à partir de connaissances a priori. Modélisation et simulation. Université de Grenoble, 2013. Français. ⟨tel-00965163v1⟩



Record views


Files downloads