The project is divided into 3 workpackages:
- WP1: Modeling surgical knowledge
- WP2: Kinematic data analysis
- WP3: Validation and evaluation
Progress:
WP1: Modeling surgical knowledge
T1.1: Surgical Process Modeling
a- Semantic models
* Ontology specific to the robot-assisted hysterectomy
– Build upon OntoSPM (core ontology for surgical process models)
– Pre-existing terms from FMA, OntoSPM, LapOntoSPM
– Knowledge elicitation (interview of 2 gynecological surgeons): additions specific to the hysterectomy procedure
* To do: relations, validation?
b- Procedural models
* generic Surgical Process Model (gSPM) at 2 levels of granularity (phases and steps)
– Knowledge elicitation (interviews of 1 gynecological surgeon): gSPM for robot-assisted hysterectomy
* Transcription of the gSPM in the xml format
T1.2: Surgical Process Learning
1- gSPM and variability
* review of articles
– study of variability
– article almost finalized for submission to journal
* gSPM formalization
– evolution of initial definition for a more compact representation
– XML definition
2- from gSPM to svGEM (in progress)
transformation of gSPM in more formal AI mathematical models
* gSPM to colored Petri nets (done)
(-) small loss of understandability
(+) simulation algorithm to generate all coherent iSPMs
– proposition of an algorithm gSPM to CPN
– implementation and validation
* gSPM to state variables Graphical event models (via CPN) (in progress)
(-) small loss of understandability
(+) better representation of variability
(+) simulation algorithm to generate all coherent iSPMs
(+) existence of svGEM learning algorithms
– proposition of an algorithm gSPM/CPN to svGEM
– validation of this algorithm
– proposition of a gSPM/svGEM learning algorithm from iSPMs (and one given gSPM) (in progress)
WP2- Kinematic data analysis
WP3- Validation and Evaluation
a- Validation
* Using experts knowledge: Delphi consensus
– Questionnaire including the vocabulary (phases, steps and their definition) and the process (gSPM)
– 5 rounds
– 35 surgical experts (> 10 years of experience in robot-assisted hysterectomy) in the 1st round, 26 surgical experts in the last round
– Article in progress
* Using experts practice: comparison between the gSPM and actual surgeries
– 31 surgeries recorded (from 6 surgeons)
– 4 comparison metrics
– Article almost finalized for submission
- Validation of gSPM to CPN conversion
- Validation of gSPM/CPN to svGEM conversion
