TY - GEN
T1 - Virtually assisted optical colonoscopy
AU - Marino, Joseph
AU - Qiu, Feng
AU - Kaufman, Arie
PY - 2008
Y1 - 2008
N2 - We present a set of tools used to enhance the optical colonoscopy procedure in a novel manner with the aim of improving both the accuracy and efficiency of this procedure. In order to better present the colon information to the gastroenterologist performing a conventional (optical) colonoscopy, we undistort the radial distortion of the fisheye view of the colonoscope. The radial distortion is modeled with a function that converts the fisheye view to the perspective view, where the shape and size of polyps can be more readily observed. The conversion, accelerated on the graphics processing unit and running in real-time, calculates the corresponding position in the fisheye view of each pixel on the perspective image. We also merge our previous work in computer-aided polyp detection for virtual colonoscopy into the optical colonoscopy environment. The physical colonoscope path in the optical colonoscopy is approximated with the hugging comer shortest path, which is correlated with the centerline in the virtual colonoscopy. With the estimated distance that the colonoscope has been inserted, we are able to provide the gastroenterologist with visual cues along the observation path as to the location of possible polyps found by the detection process. In order to present the information to the gastroenterologist in a non-intrusive manner, we have developed a friendly user interface to enhance the optical colonoscopy without being cumbersome, distracting, or resulting in a more lackadaisical inspection by the gastroenterologist.
AB - We present a set of tools used to enhance the optical colonoscopy procedure in a novel manner with the aim of improving both the accuracy and efficiency of this procedure. In order to better present the colon information to the gastroenterologist performing a conventional (optical) colonoscopy, we undistort the radial distortion of the fisheye view of the colonoscope. The radial distortion is modeled with a function that converts the fisheye view to the perspective view, where the shape and size of polyps can be more readily observed. The conversion, accelerated on the graphics processing unit and running in real-time, calculates the corresponding position in the fisheye view of each pixel on the perspective image. We also merge our previous work in computer-aided polyp detection for virtual colonoscopy into the optical colonoscopy environment. The physical colonoscope path in the optical colonoscopy is approximated with the hugging comer shortest path, which is correlated with the centerline in the virtual colonoscopy. With the estimated distance that the colonoscope has been inserted, we are able to provide the gastroenterologist with visual cues along the observation path as to the location of possible polyps found by the detection process. In order to present the information to the gastroenterologist in a non-intrusive manner, we have developed a friendly user interface to enhance the optical colonoscopy without being cumbersome, distracting, or resulting in a more lackadaisical inspection by the gastroenterologist.
KW - Centerline extraction
KW - Distortion correction
KW - Optical colonoscopy
KW - Radial undistortion
KW - Shortest path
KW - Virtual colonoscopy
UR - https://www.scopus.com/pages/publications/44349147128
U2 - 10.1117/12.770864
DO - 10.1117/12.770864
M3 - Conference contribution
SN - 9780819471000
T3 - Progress in Biomedical Optics and Imaging - Proceedings of SPIE
BT - Medical Imaging 2008 - Physiology, Function, and Structure from Medical Images
T2 - Medical Imaging 2008 - Physiology, Function, and Structure from Medical Images
Y2 - 17 February 2008 through 19 February 2008
ER -