Robert DiPietro

Johns Hopkins University, Amazon (United States)

Papers

Total Citations

H-Index

About

Robert DiPietro is a leading researcher at the intersection of surgical robotics and machine perception, with key contributions spanning surgical activity recognition, recurrent neural network (RNN) architectures, and robust visual-inertial SLAM. His most influential work, "Segmenting and classifying activities in robot-assisted surgery with recurrent neural networks" (2019, 73 citations), pioneered deep learning methods for parsing complex surgical workflows from kinematic and video data, directly enabling automated skill assessment and intraoperative decision support. Earlier foundational work on recognizing surgical activities with RNNs (2016, 20 citations) established temporal modeling as a cornerstone of surgical data science. More recently, DiPietro has tackled the critical challenge of failure detection and recovery in large-scale indoor SLAM (2024, 5 citations), integrating camera and IMU data to create navigation systems resilient to real-world drift and occlusion—a vital advance for autonomous robotics in hospitals, factories, and warehouses. His research uniquely bridges the gap between high-stakes surgical applications and general robotic perception, demonstrating how deep sequential models can transform both patient care and autonomous navigation. With a growing citation footprint and a trajectory toward robust, real-world deployment, DiPietro is shaping the future of context-aware robotic systems.