Autonomous surgical lamps

Phanaphon Sereelertwiwat (Kim)
advisor: Jörn Teuber

Autonomous Surgical Lamps, originally is the project of the PhD student, Jörn Teuber. The basic idea of this project is to adjust the surgical lamps to their optimal positions in open surgeries with the use of a depth camera trackers. This data are tested in a VR simulation to ensure the outcome of the algorithm. It is implemented on OpenCV under C++ language. This project primitively has been working on Linux operating system, however, due to my insufficient capacity of my computer, I contributed this project under Windows operating system. The task I was assigned to handle is a small part of this project called as the “Gaze Tracker”. It is the method used to verify whether the surgeons are looking at the target points. Based on the records of the real situation from operation room, we extract the point cloud data from the kinect installed above the surgeons’ heads. With this data, I analyzed the heads of the surgeons based on the highest points shown in the projection. I’ve tried many approaches and the method I’ve used was to threshold the head by 20 cm from the center or the highest points. Then I contoured the remaining data to roughly drawing the person’s head. As a consequences, I could identify the major and minor axis of the heads to see direction the surgeons are facing to. I extended the line of sight from the course and as a result, I could calculate whether the surgeons are focusing at the target position or not within total of 30 degree angle from estimated from the center of surgeons’ head.