Vision-servo system for middle ear treatment

Round window, the boundary between the middle and inner ear, plays a well-established role in cochlear physiology. It is usually inaccessible to localized medical treatment. In this paper, a novel approach is presented to explore local therapy of round window, leveraging on an intelligent vision component for visualizing and locating the round window, and a servo controlled working channel to administer drug implant placement. Deviating from the current and conventional methods of carrying out these operations, the proposed approach can be automated with high precision and control, resulting in efficient and less invasive execution of such middle ear treatment. The vision-servo system incorporates intelligent algorithms to carry out Haar-like features detection, image processing and PID motion control. When the round window is detected via computer vision, the working channel (needle) carrying medication is controlled by a precise motion controller with feedback from the vision component to move the dispensing tip to the target point automatically. Compared to traditional approaches, the proposed one is more efficient, employing precise sensing and manipulation technology outperforming the human senses and operations, and thus reducing invasiveness, enhancing wound healing rate and reducing the risk of infections. Pre-clinical experiments are carried out to verify the feasibility and efficacy of the proposed approach.