Computerized detection of lumen boundary in robotic colonoscopy for medical diagnosis

Abstract

A Colonoscope is an endoscopic instrument used by physicians to view the inside of the colon for diagnosing possible colonic diseases. In difficult cases of diagnostic colonoscopy, the tip of the endoscope may not be properly maneuverable and controllable. This could result in a painful situation for the patient and perforations in intestinal wall may occur, leading to undesirable consequences. A new method is being considered to provide assistance in solving such problems by employing an intelligent computer vision system in colonoscopy, thus enabling automatic identification of the lumen contour and deciding on the direction to proceed. This approach will provide visual feedback and guidance to the endoscopist, forming the basis for the new method of computerized lumen boundary detection from colonoscopic images. Moreover, estimating the properties of lumen by analyzing the colonoscopic images would help in the identification of lesions and abnormal tissues at an early stage.