A novel approach to telemedicine video control technology

Telemedicine, especially teletrauma and telepresence, relies on live video streams, for optimal patient diagnosis and management, including assessment of diagnostic images such as plain x-ray, ultrasound, or other modalities. A pan-tilt-zoom camera used in such cases should provide enough freedom of movement and a real sense of telepresence. Operating such cameras, especially in a critical situation, is time consuming, having to control multiple parameters one at a time to get the desired view on patient. In this article, we propose a novel approach to remote camera control for use in telemedicine. A method using a neural network for visual servo control is proposed. The system is based on two cameras: one for live visual feedback and the other for visual servoing of the robot arm to the desired location. The operator selects the location he wants to observe and the visual servo navigates a live feedback camera accordingly.