Data & Code: Diver Pose Estimation

The Constructor Robotics group contributed within the EU-project “Cognitive autonomous diving buddy (CADDY)” to the development of methods for the machine perception of divers (see, e.g., [2] [4]) in the context of Underwater Human Machine Interaction (U-HRI) [1]. This included among others the collection of data for diver pose estimation.

The CADDY Underwater Diver Pose Dataset [3] consists of 12,000 stereo pair images synchronized with diver body pose measurements from a suit of Inertial Measurement Units (IMUs) on the diver’s body called DiverNet [3].


[1] A. Birk, “A Survey of Underwater Human-Robot Interaction (U-HRI),” Current Robotics Reports, Springer Nature, vol. 3, pp. 199-211, 2022. [Open Access]

[2] A. G. Chavez, A. Ranieri, D. Chiarella, and A. Birk, “Underwater Vision-Based Gesture Recognition: A Robustness Validation for Safe Human-Robot Interaction,” IEEE Robotics and Automation Magazine (RAM), vol. 28, pp. 67-78, 2021. [Preprint]

[3] A. G. Chavez, A. Ranieri, D. Chiarella, E. Zereik, A. Babic, and A. Birk, “CADDY Underwater Stereo-Vision Dataset for Human-Robot Interaction (HRI) in the Context of Diver Activities,” Journal of Marine Science and Engineering (JMSE), spec.iss. Underwater Imaging, vol. 7, 2019. [Open Access]

[4] A. G. Chavez, C. A. Mueller, A. Birk, A. Babic, and N. Miskovic, “Stereo-vision based diver pose estimation using LSTM recurrent neural networks for AUV navigation guidance,” in IEEE Oceans, Aberdeen, UK, 2017. [Preprint]