Motivation: Treatment of Stroke Patients ---------- In emergency situations, body weight adapted dosage is crucial for many drugs. This is particularly important if drugs are known to have a narrow therapeutic range with decreased efficacy in lower dosages and an increased risk for possibly severe adverse effects in higher dosages. Many emergency patients are unable to communicate information on their body weight because of their symptoms, e.g. decreased consciousness or neurological disorders, or because they simply do not know their own body weight. In addition, severe injuries or motor symptoms prohibit easy weighing procedures for many patients. Certain diseases like ischemic stroke are associated with a very narrow time window for treatment and do not allow time to weigh each patient in the emergency situation. Therefore, visual estimation of the patient’s body weight by the attending physician in the emergency room has become routine worldwide. This approach bears the risk of estimation errors and may result in dosing errors, which has been shown for weight- based emergency medications. Less complicated and more precise methods to evaluate body weight are required for emergency patients to minimize potential dosing errors. The time required to evaluate body weight should be as short as possible and new methods should be easy to integrate into the practical processes of an emergency room to avoid treatment delays. ![Scene from trauma room with sensors integrated in the ceiling. ][1] Libra3D was designed with the goal of creating a new contact-free method to evaluate body weight based on visual sensor technology. Furthermore, feasibility and accuracy of this method were also evaluated. The system uses a 3D sensors, e.g. Microsoft Kinect, Microsoft Kinect One, as well as a thermal camera for easier segmentation. Processing ---------- -- coming soon -- Datasets -------- The datasets for testing and training of the applied model are made public here in OSF. Due to privacy issues the color data is deleted from the files. The data is stored in the common PCL file format, containing Cartesian points, as well as the thermal data, if available. The filename containts the metadata: Gender_GroundTruthWeight_ID_SensorType_0000.pcd The **Gender** can be F or M. The **GroundTruthWeight** is given in kilograms. The **ID** is an arbitrary value. Via **SensorType** can be K1 for Microsoft Kinect, or K2 for the Kinect One. Please contact me for access to the available datasets. Publications ------------ Within the project Libra3D several papers are produced. If you refere to this dataset please cite the latest publication. - Christian Pfitzner, Stefan May and Andreas Nüchter: **Evaluation of Features from RGB-D Data for Human Body Weight Estimation**, *In Proceedings of the 20th World Congress of the International Federation of Automatic Control (IFAC)*, 9-14 July 2017, Toulouse, France, (accepted) - Christian Pfitzner, Stefan May, Christian Merkl: Vorrichtung und **Verfahren zur optischen Erfassung eines Gewichtes einer Person**, German Patent. Submission 29. Februar 2016. - Christian Pfitzner, Stefan May, and Andreas Nüchter: **Neural Network-based Visual Body Weight Estimation for Drug Dosage Finding**. *In Proceedings of the SPIE Medical Imaging Conference on Image Processing*, San Diego, CA, USA, March 2016. - Christian Pfitzner, Stefan May, Christian Merkl, Lorenz Breuer, Martin Köhrmann, Joel Braun, Franz Dirauf, and Andreas Nüchter. **Libra3D: Body Weight Estimation for Emergency Patients in Clinical Environment with a 3D Structured Light Sensor**, *In Proceedings of the IEEE International Conference Robotics and Automation (ICRA '15)*, Seattle, WA, USA, May 2015. - Christian Pfitzner. **Robotic Vision in Medical Applications: Visual Weight Estimation for Emergency Patients**. *Workshop Proposal to the Ph.D. Forum at the International Conference Robotics and Automation (ICRA '15)*, Seattle, WA, USA, May 2015. [1]: https://mfr.osf.io/export?url=https://osf.io/84j6y/?action=download&direct&mode=render&initialWidth=848&childId=mfrIframe&format=1200x1200.jpeg