Image Based High-Level Control System Design for Steering and Controlling of an Active Capsule Endoscope

Abstract

Using wireless capsule endoscopes for diagnosis and treatment of gastrointestinal tract disease has been increased in recent years. Due to the anatomical shape of the stomach, one of the most important challenges in capsule endoscopy is steering and controlling of the capsule inside the stomach to obtain the high-quality imaging. Another challenge is improving the accuracy of steering and controlling of the capsule, which is already highly dependent upon the operator's knowledge and skill who acts as a high-level controller. To address these challenges, in this paper, a novel image based High-Level Control system (H-LC) is proposed for real-time autonomous control of an active capsule endoscope. In the proposed system, first, a new desired path planning method based on detecting the stomach folds is implemented using the acquired images. Then, various maneuvers are presented for following the desired path and adjusting the orientation of the capsule. The core of these maneuvers is based on geometric analysis and computer vision techniques. The results for path planning method show the averages of Sensitivity 94, Specificity 88 and Accuracy 92 on 82 endoscopic images of the stomach. For evaluating the proposed maneuver planning method, simulation in MATLAB is used. Based on the results, the suggested H-LC system can be used accurately for planning and following the desired path to achieve a more accurate control and steering of a capsule endoscope in a stomach with an unknown size.