TY - JOUR
T1 - Comparative Validation of Polyp Detection Methods in Video Colonoscopy
T2 - Results from the MICCAI 2015 Endoscopic Vision Challenge
AU - Bernal, Jorge
AU - Tajkbaksh, Nima
AU - Sanchez, Francisco Javier
AU - Matuszewski, Bogdan J.
AU - Chen, Hao
AU - Yu, Lequan
AU - Angermann, Quentin
AU - Romain, Olivier
AU - Rustad, Bjorn
AU - Balasingham, Ilangko
AU - Pogorelov, Konstantin
AU - Choi, Sungbin
AU - Debard, Quentin
AU - Maier-Hein, Lena
AU - Speidel, Stefanie
AU - Stoyanov, Danail
AU - Brandao, Patrick
AU - Cordova, Henry
AU - Sanchez-Montes, Cristina
AU - Gurudu, Suryakanth R.
AU - Fernandez-Esparrach, Gloria
AU - Dray, Xavier
AU - Liang, Jianming
AU - Histace, Aymeric
N1 - Funding Information:
This work was supported in part by ASU-Mayo Clinic partnerships, in part by the Spanish Government through the Funded Project iVENDIS under Project DPI2015-65286-R, in part by FSEED, in part by the Secretaria d'Universitats i Recerca de la Generalitat de Catalunya under Grant 2014-SGR-1470 and Grant 2014-SGR-135, in part by par SATT IdFInnov (France) through the Project Smart Videocolonoscopy under Grant 186, and in part by the European Union through the ERC starting grant COMBIOSCOPY under the New Horizon Framework Programme under Grant ERC-2015-StG-37960. The authors would like to thank EndoVis challenge organizers for their continuous help and guidance through both challenge and paper preparation. The idea of organizing a competition for polyp detection in colonoscopy was first conceived by Dr. Jianming Liang, and the foundational framework was established by Drs. Tajbakhsh and Liang before the first challenge at ISBI-2015. The associated ground truth images in the ASU-Mayo Clinic Colonoscopy Video c? Database were created by Saiswathi Javangula, Ireen Khan, Kamran Bodushev, Sarah Fallah-Adl, and Tracy Phan. The ASU-Mayo Clinic Colonoscopy Video ? Database is copyrighted and its use is granted to the work for the challenge on polyp detection in colonoscopy as reported this IEEE TMI paper. For any other uses, a prior agreement must be obtained from Arizona State University.
Publisher Copyright:
© 1982-2012 IEEE.
PY - 2017/6
Y1 - 2017/6
N2 - Colonoscopy is the gold standard for colon cancer screening though some polyps are still missed, thus preventing early disease detection and treatment. Several computational systems have been proposed to assist polyp detection during colonoscopy but so far without consistent evaluation. The lack of publicly available annotated databases has made it difficult to compare methods and to assess if they achieve performance levels acceptable for clinical use. The Automatic Polyp Detection sub-challenge, conducted as part of the Endoscopic Vision Challenge (http://endovis.grand-challenge.org) at the international conference on Medical Image Computing and Computer Assisted Intervention (MICCAI) in 2015, was an effort to address this need. In this paper, we report the results of this comparative evaluation of polyp detection methods, as well as describe additional experiments to further explore differences between methods. We define performance metrics and provide evaluation databases that allow comparison of multiple methodologies. Results show that convolutional neural networks are the state of the art. Nevertheless, it is also demonstrated that combining different methodologies can lead to an improved overall performance.
AB - Colonoscopy is the gold standard for colon cancer screening though some polyps are still missed, thus preventing early disease detection and treatment. Several computational systems have been proposed to assist polyp detection during colonoscopy but so far without consistent evaluation. The lack of publicly available annotated databases has made it difficult to compare methods and to assess if they achieve performance levels acceptable for clinical use. The Automatic Polyp Detection sub-challenge, conducted as part of the Endoscopic Vision Challenge (http://endovis.grand-challenge.org) at the international conference on Medical Image Computing and Computer Assisted Intervention (MICCAI) in 2015, was an effort to address this need. In this paper, we report the results of this comparative evaluation of polyp detection methods, as well as describe additional experiments to further explore differences between methods. We define performance metrics and provide evaluation databases that allow comparison of multiple methodologies. Results show that convolutional neural networks are the state of the art. Nevertheless, it is also demonstrated that combining different methodologies can lead to an improved overall performance.
KW - Endoscopic vision
KW - handcrafted features
KW - machine learning
KW - polyp detection
KW - validation framework
UR - http://www.scopus.com/inward/record.url?scp=85021449496&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85021449496&partnerID=8YFLogxK
U2 - 10.1109/TMI.2017.2664042
DO - 10.1109/TMI.2017.2664042
M3 - Article
C2 - 28182555
AN - SCOPUS:85021449496
VL - 36
SP - 1231
EP - 1249
JO - IEEE Transactions on Medical Imaging
JF - IEEE Transactions on Medical Imaging
SN - 0278-0062
IS - 6
M1 - 7840040
ER -