Colonoscopy is the most-commonly adopted visual screening procedure of the colon by means of a flexible tiny endoscopic camera. In an effort to promote early screening of colorectal cancer, and to facilitate mastering the endoscope motion by the physician, teleoperable robotic endoscopes and computer-assisted colonoscopy systems are being developed.
Our goal is to devise computer-based methods to analyze the endoscope video stream in an effort to support the endoscopist's inspection of the colon mucosa.
We are researching advanced monocular endoscope localization strategies that can cope with the disruptive events of colonoscopy videos, such as mucosa deformations, lack of texture, and rapid camera motions. Localization is a key component of closed-loop control of teleoperable robotic devices. In particular, we have presented a first extensive comparison of state-of-the-art vision-based localization methods for flexible endoscopes.
In particular, we compared (see Fig.1) the performance of two state-of-the-art 6 degrees of freedom (DoF) ego-motion estimation algorithms, namely, Artificial Neural Networks (ANN) and Visual Odometry (VO), under four optical-flow (OF) algorithms. The ability for each of these methods to precisely localize the camera after a long trajectory have been examined.
Figure 1: Comparison of the (ending point) localization error for supervised and unsupervised monocular localization methods (for different optical-flow algorithms).
Figure 2: Localization in a synthetic 3-D virtual model of a colon.
Figure 3: Localization in a synthetic 3-D virtual model of a colon.
G. A. Puerto-Souza, A. Staranowicz, C. Bell, P. Valdastri, and G.L. Mariottini, "A Comparative Study of Ego-Motion Estimation Algorithms for Teleoperated Robotic Endoscope", Medical Image Computing and Computer Assisted Interventions (MICCAI'14) - CARE Workshop, Boston, 2014.
C. Bell, G. A. Puerto-Souza, G.L. Mariottini and P. Valdastri, "Six DOF Motion Estimation for Teleoperated Flexible Endoscopes Using Optical Flow: A Comparative Study", 2014 IEEE International Conference on Robotics and Automation, Hong Kong, China, 2014.
File translated from
version 3.85. On 28 Oct 2014, 22:52.