The Evolino-trained LSTM recurrent neural network successfully automated suture knot winding, reducing the total time for the entire knot to 25.8 seconds compared to 33.7 seconds for the preprogrammed controller.
Using recurrent neural networks trained with the Evolino algorithm enables robotic surgical systems to learn and automate complex tasks like suture knot tying from human demonstrations.
Absolute Event Rate: 25.8% vs 33.7%
Tying suture knots is a time-consuming task performed frequently during minimally invasive surgery (MIS). Automating this task could greatly reduce total surgery time for patients. Current solutions to this problem replay manually programmed trajectories, but a more general and robust approach is to use supervised machine learning to smooth surgeon-given training trajectories and generalize from them. Since knot tying generally requires a controller with internal memory to distinguish between identical inputs that require different actions at different points along a trajectory, it would be impossible to teach the system using traditional feedforward neural nets or support vector machines. Instead we exploit more powerful, recurrent neural networks (RNNs) with adaptive internal states. Results obtained using long short-term memory RNNs trained by the recent Evolino algorithm show that this approach can significantly increase the efficiency of suture knot tying in MIS over preprogrammed control.
Mayer et al. (Tue,) conducted a other in Suture knot tying in minimally invasive surgery. Evolino-trained LSTM recurrent neural network vs. Preprogrammed controller was evaluated on Total time for the entire knot (seconds). The Evolino-trained LSTM recurrent neural network successfully automated suture knot winding, reducing the total time for the entire knot to 25.8 seconds compared to 33.7 seconds for the preprogrammed controller.