Surgical Instrument Signaling Gesture Recognition Using Surface Electromyography Signals

Area of Science:

• Biomedical Engineering
• Human-Computer Interaction
• Robotics

Background:

• Surgical Instrument Signaling (SIS) relies on hand gestures for surgeon-instrumentator communication, crucial for preventing errors.
• Existing SIS systems often use limited gesture sets, potentially hindering nuanced communication.
• Wearable technology offers a promising avenue for real-time, non-invasive surgical communication systems.

Purpose of the Study:

• To assess the feasibility of a gesture recognition system for SIS using surface electromyographic (sEMG) signals.
• To develop and evaluate a processing routine for wearable SIS applications, particularly for telesurgery and robotic surgery.
• To create a comprehensive database of 14 SIS gestures from 10 volunteers for robust system training and testing.

Main Methods:

• Acquisition of sEMG signals from the Myo armband during the execution of 14 distinct SIS gestures by 10 volunteers.
• Implementation of a processing routine involving automatic segmentation, feature extraction (13 sets), and feature selection.
• Classification of sEMG signals using 6 different classifiers, including Support Vector Machine (SVM), and 2 ensemble techniques.

Main Results:

• An accuracy of 76% was achieved using the SVM classifier across all recorded databases.
• Individual volunteer analysis yielded a higher accuracy of 88% with the SVM classifier.
• The developed system demonstrated suitability for real-time SIS gesture recognition in wearable applications.

Conclusions:

• Surface electromyography (sEMG) signals are effective for recognizing a diverse set of Surgical Instrument Signaling (SIS) gestures.
• The proposed processing routine and classification methods are suitable for wearable SIS applications, enhancing surgical communication.
• This research provides a foundation for integrating advanced gesture recognition into telesurgery and robotic surgery systems.