Unsupervised Anomaly Detection for a Smart Autonomous Robotic Assistant Surgeon (SARAS) Using a Deep Residual Autoencoder

Anomaly detection in Minimally-Invasive Surgery (MIS) traditionally requires a human expert monitoring the procedure from a console, whereas automated anomaly detection systems in this area typically rely on classical supervised learning. Anomalous surgical events, however, are rare, making it difficult to capture data to train a model in a supervised fashion. In this work we propose an unsupervised approach to anomaly detection for robotic MIS based on deep residual autoencoders. The idea is to make the autoencoder learn the `normal' distribution of the data and detect abnormal events deviating from this distribution by measuring a reconstruction error. The model is trained and validated upon both the publicly available Cholec80 dataset and a set of videos captured on procedures using artificial anatomies (`phantoms') as part of the Smart Autonomous Robotic Assistant Surgeon (SARAS) project. The system achieves recall and precision equal to 78.4%, 91.5%, respectively, on Cholec80 and of 95.6%, 88.1% on the SARAS phantom dataset. The system was developed and deployed as part of the SARAS platform for real-time anomaly detection with a processing time of 25 ms per frame.