TY - GEN
T1 - Auto-calibration Method for Active 3D Endoscope System Using Silhouette of Pattern Projector
AU - Furukawa, Ryo
AU - Naito, Masahito
AU - Miyazaki, Daisuke
AU - Baba, Masahi
AU - Hiura, Shinsaku
AU - Sanomura, Yoji
AU - Tanaka, Shinji
AU - Kawasaki, Hiroshi
N1 - Funding Information:
Acknowledgment. This work was supported in part by JSPS KAKENHI Grant No. 15H02779, 16H02849, MIC SCOPE 171507010 and MSR CORE12.
Publisher Copyright:
© 2018, Springer International Publishing AG, part of Springer Nature.
PY - 2018
Y1 - 2018
N2 - In this paper, we develop an active stereo system for endoscope which requires auto-calibration, because a micro pattern projector is inserted through the instrument channel during an operation and cannot be fixed to the endoscope. For solution, a new auto-calibration technique with full 6-DOF estimation of an active stereo system without any extra devices nor extra pattern projections is proposed. In the technique, the pattern projector itself is simultaneously captured with a target scene by an endoscope camera and the silhouette of the pattern projector is used to conduct 2D-3D matching by using the knowledge of the shape of the projector. In addition, the markers which is included in the projection pattern are extracted and the distances from the closest epipolar lines are calculated as for the cost function. To enhance the robustness of the reconstruction, we also propose a simple high dynamic range (HDR) imaging system for an endoscope by alternating the input power of the pattern projector ON and OFF to blink the pattern so that exposure time will vary with beat frequency, realizing a virtual multi-exposure camera. By applying our auto calibration technique with HDR imaging system, we achieved a robust and accurate reconstruction of tissue in metric 3D under practical operation of the endoscopic system, such as reconstruction of the inside of a real stomach of a pig.
AB - In this paper, we develop an active stereo system for endoscope which requires auto-calibration, because a micro pattern projector is inserted through the instrument channel during an operation and cannot be fixed to the endoscope. For solution, a new auto-calibration technique with full 6-DOF estimation of an active stereo system without any extra devices nor extra pattern projections is proposed. In the technique, the pattern projector itself is simultaneously captured with a target scene by an endoscope camera and the silhouette of the pattern projector is used to conduct 2D-3D matching by using the knowledge of the shape of the projector. In addition, the markers which is included in the projection pattern are extracted and the distances from the closest epipolar lines are calculated as for the cost function. To enhance the robustness of the reconstruction, we also propose a simple high dynamic range (HDR) imaging system for an endoscope by alternating the input power of the pattern projector ON and OFF to blink the pattern so that exposure time will vary with beat frequency, realizing a virtual multi-exposure camera. By applying our auto calibration technique with HDR imaging system, we achieved a robust and accurate reconstruction of tissue in metric 3D under practical operation of the endoscopic system, such as reconstruction of the inside of a real stomach of a pig.
UR - http://www.scopus.com/inward/record.url?scp=85042495712&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85042495712&partnerID=8YFLogxK
U2 - 10.1007/978-3-319-75786-5_19
DO - 10.1007/978-3-319-75786-5_19
M3 - Conference contribution
AN - SCOPUS:85042495712
SN - 9783319757858
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 222
EP - 236
BT - Image and Video Technology - 8th Pacific-Rim Symposium, PSIVT 2017, Revised Selected Papers
A2 - Hitoshi, Carlos
A2 - Paul, Manoranjan
A2 - Huang, Qingming
PB - Springer Verlag
T2 - 8th Pacific-Rim Symposium on Image and Video Technology, PSIVT 2017
Y2 - 20 November 2017 through 24 November 2017
ER -