Fig. 3
From: Artificial intelligence for phase recognition in complex laparoscopic cholecystectomy
Accuracy of the MS-TCN and the inter-rater agreement score stratified by LC complexity level. a For the simplest LC procedures, the MS-TCN reaches 92% accuracy, similar to the agreement score between the annotators. As expected, as the complexity level increases, the accuracy of the MS-TCN model decreases. Nevertheless, for the most complex LC procedures (levels 4 and 5), the MS-TCN model reaches 81% accuracy. b Results stratified by institution; each marker in the graph represents a different hospital source. The x-value of each marker is the average complexity level of LC procedures for the given source hospital. The y-value of each marker is the average accuracy achieved by the AI model on LC procedures for the given source hospital