Stereo vision based intelligent system for middle ear treatment

Round window, a channel connecting the middle and inner ears, plays an important role in hearing process of human beings. Usually, localized medical treatment to the round window requires surgical treatment to be carried out by professional surgeons. In this paper, we propose a novel approach to explore the potential of robotic-based round window local therapy, leveraging on the technique of stereo computer vision and precise motion control algorithm. In our design, an endoscope acquires stereo images to reconstruct the 3D structure of the middle ear for locating the round window, and a robotic-based working channel will administer the process of drug placement. Once the round window is located via computer vision, the working channel with its end effector (dispensing needle) controlled by a motion controller will be driven to the target point under constrictive conditions of the middle ear. Compared to the traditional manual approaches, the proposed system can render a more accurate and efficient performance, employing precise sensing and manipulation technology outperforming the human senses and control, and thus reducing invasiveness and enhancing the surgery safety. Pre-clinical experiments are carried out to validate the feasibility of the proposed approach.