Open-H-Embodiment Dataset for Medical Robotics Foundation Models
To tackle the lack of data in autonomous medical robotics, a new dataset named Open-H-Embodiment has been launched. This dataset is the most extensive open collection of medical robotic videos with synchronized kinematics, encompassing over 49 institutions and various robotic systems such as CMR Versius, da Vinci from Intuitive Surgical, da Vinci Research Kit (dVRK), Rob Surgical BiTrack, MIRA from Virtual Incision, Moon Surgical Maestro, and custom setups. It includes data on surgical manipulation, robotic ultrasound, and endoscopy. Using this dataset, two foundational models have been showcased, with the goal of enhancing autonomous medical robotics through improved AI training data.
Key facts
- Open-H-Embodiment is the largest open dataset of medical robotic video with synchronized kinematics.
- Dataset spans over 49 institutions.
- Includes multiple robotic platforms: CMR Versius, Intuitive Surgical's da Vinci, da Vinci Research Kit (dVRK), Rob Surgical BiTrack, Virtual Incision's MIRA, Moon Surgical Maestro, and custom systems.
- Covers surgical manipulation, robotic ultrasound, and endoscopy procedures.
- Two foundation models have been demonstrated using the dataset.
- Existing medical robotic datasets are small, single-embodiment, and rarely shared openly.
- Autonomous medical robots aim to improve patient outcomes, reduce provider workload, democratize access to care, and enable superhuman precision.
- The dataset is introduced in arXiv paper 2604.21017.
Entities
Institutions
- CMR Versius
- Intuitive Surgical
- da Vinci Research Kit (dVRK)
- Rob Surgical BiTrack
- Virtual Incision
- MIRA
- Moon Surgical Maestro