Robotic-Assisted Automatic Orientation and Insertion for Bronchoscopy Based on Image Guidance

Abstract

Due to the large number and complex structure of bronchial branches, frequent orientation and insertions during bronchoscopy are likely to cause surgeon fatigue and errors. Therefore, we designed a robotic-assisted automatic intervention system (RAIS) based on image guidance, which aims to realize automatic orientation and insertion of bronchoscope, for improving the intelligence and efficiency of the bronchoscopy. To realize image guidance, we proposed a highly robust and accurate lumen center detection method, which combines deep learning-based object detection and histogram back-projection. With image guidance of lumen center, RAIS automatically completes the orientation and insertion of bronchoscope. The results of human phantom lung experiments show that the accuracy and recall of proposed lumen center method can respectively reach 94.12% and 86.23% when the bronchoscope is being localized by RAIS. Moreover, the results show that it is feasible to automate the orientation and insertion of bronchoscope using RAIS, which is more efficient than the manual operating the robot. The proposed RAIS could replace the surgeon in performing a large number of repetitive and simple tasks, which can efficiently reduce the workload of the surgeon and improve the safety and efficiency of the procedure.