Enhancing normal-abnormal classification accuracy in colonoscopy videos via temporal consistency

Gustavo A. Puerto-Souza; Siyamalan Manivannan; María P. Trujillo; Jesus A. Hoyos; Emanuele Trucco; Gian-Luca Mariottini

doi:10.1007/978-3-319-29965-5_13

Enhancing normal-abnormal classification accuracy in colonoscopy videos via temporal consistency

Gustavo A. Puerto-Souza (Lead / Corresponding author), Siyamalan Manivannan, María P. Trujillo, Jesus A. Hoyos, Emanuele Trucco, Gian-Luca Mariottini

Computing

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

1 Citation (Scopus)

Abstract

This paper proposes a novel hierarchical approach to improve the accuracy of the classification of normal-vs-abnormal frames in white-light colonoscopy videos. The existing approaches label each frame independently, without considering the temporal consistency between adjacent frames. Temporal consistency, however, can improve the classification accuracy in the presence of unclear/uncertain images. We propose to leverage temporal consistency between adjacent frames for colonoscopy video frame classification using a novel hierarchical classifier. Comparative experiments with five challenging full colonoscopy videos show that the proposed approach considerably improves the mean class normal/abnormal classification accuracy compared to the approaches where the frames are classified independently.

Original language	English
Title of host publication	Computer-Assisted and Robotic Endoscopy
Subtitle of host publication	Second International Workshop, CARE 2015, Held in Conjunction with MICCAI 2015, Munich, Germany, October 5, 2015, Revised Selected Papers
Editors	Xiongbiao Luo, Tobias Reichl, Austin Reiter , Gian-Luca Mariottini
Publisher	Springer Verlag
Pages	129-139
Number of pages	11
Volume	9515
ISBN (Electronic)	9783319299655
ISBN (Print)	9783319299648
DOIs	https://doi.org/10.1007/978-3-319-29965-5_13
Publication status	Published - 20 Feb 2016
Event	2nd International Workshop on Computer-Assisted and Robotic Endoscopy: CARE 2015 - Munich, Germany Duration: 5 Oct 2015 → 5 Oct 2015 http://ranger.uta.edu/~gianluca/CARE15/Main.html (Link to conference)

Publication series

Name	Lecture notes in computer science
Volume	9515
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Workshop

Workshop	2nd International Workshop on Computer-Assisted and Robotic Endoscopy
Country/Territory	Germany
City	Munich
Period	5/10/15 → 5/10/15
Internet address	http://ranger.uta.edu/~gianluca/CARE15/Main.html (Link to conference)

ASJC Scopus subject areas

General Computer Science
Theoretical Computer Science

Access to Document

10.1007/978-3-319-29965-5_13

Cite this

Puerto-Souza, G. A., Manivannan, S., Trujillo, M. P., Hoyos, J. A., Trucco, E., & Mariottini, G.-L. (2016). Enhancing normal-abnormal classification accuracy in colonoscopy videos via temporal consistency. In X. Luo, T. Reichl, A. Reiter , & G.-L. Mariottini (Eds.), Computer-Assisted and Robotic Endoscopy: Second International Workshop, CARE 2015, Held in Conjunction with MICCAI 2015, Munich, Germany, October 5, 2015, Revised Selected Papers (Vol. 9515, pp. 129-139). (Lecture notes in computer science ; Vol. 9515). Springer Verlag. https://doi.org/10.1007/978-3-319-29965-5_13

Puerto-Souza, Gustavo A. ; Manivannan, Siyamalan ; Trujillo, María P. et al. / Enhancing normal-abnormal classification accuracy in colonoscopy videos via temporal consistency. Computer-Assisted and Robotic Endoscopy: Second International Workshop, CARE 2015, Held in Conjunction with MICCAI 2015, Munich, Germany, October 5, 2015, Revised Selected Papers. editor / Xiongbiao Luo ; Tobias Reichl ; Austin Reiter ; Gian-Luca Mariottini . Vol. 9515 Springer Verlag, 2016. pp. 129-139 (Lecture notes in computer science ).

@inproceedings{a3cb7151e9bc485f8f4daf87c3c72527,

title = "Enhancing normal-abnormal classification accuracy in colonoscopy videos via temporal consistency",

abstract = "This paper proposes a novel hierarchical approach to improve the accuracy of the classification of normal-vs-abnormal frames in white-light colonoscopy videos. The existing approaches label each frame independently, without considering the temporal consistency between adjacent frames. Temporal consistency, however, can improve the classification accuracy in the presence of unclear/uncertain images. We propose to leverage temporal consistency between adjacent frames for colonoscopy video frame classification using a novel hierarchical classifier. Comparative experiments with five challenging full colonoscopy videos show that the proposed approach considerably improves the mean class normal/abnormal classification accuracy compared to the approaches where the frames are classified independently.",

author = "Puerto-Souza, {Gustavo A.} and Siyamalan Manivannan and Trujillo, {Mar{\'i}a P.} and Hoyos, {Jesus A.} and Emanuele Trucco and Gian-Luca Mariottini",

year = "2016",

month = feb,

day = "20",

doi = "10.1007/978-3-319-29965-5_13",

language = "English",

isbn = "9783319299648",

volume = "9515",

series = "Lecture notes in computer science ",

publisher = "Springer Verlag",

pages = "129--139",

editor = "Luo, {Xiongbiao } and Reichl, {Tobias } and {Reiter }, {Austin } and {Mariottini }, {Gian-Luca }",

booktitle = "Computer-Assisted and Robotic Endoscopy",

address = "Germany",

note = "2nd International Workshop on Computer-Assisted and Robotic Endoscopy : CARE 2015 ; Conference date: 05-10-2015 Through 05-10-2015",

url = "http://ranger.uta.edu/~gianluca/CARE15/Main.html",

}

Puerto-Souza, GA, Manivannan, S, Trujillo, MP, Hoyos, JA, Trucco, E & Mariottini, G-L 2016, Enhancing normal-abnormal classification accuracy in colonoscopy videos via temporal consistency. in X Luo, T Reichl, A Reiter & G-L Mariottini (eds), Computer-Assisted and Robotic Endoscopy: Second International Workshop, CARE 2015, Held in Conjunction with MICCAI 2015, Munich, Germany, October 5, 2015, Revised Selected Papers. vol. 9515, Lecture notes in computer science , vol. 9515, Springer Verlag, pp. 129-139, 2nd International Workshop on Computer-Assisted and Robotic Endoscopy, Munich, Germany, 5/10/15. https://doi.org/10.1007/978-3-319-29965-5_13

Enhancing normal-abnormal classification accuracy in colonoscopy videos via temporal consistency. / Puerto-Souza, Gustavo A. (Lead / Corresponding author); Manivannan, Siyamalan; Trujillo, María P. et al.
Computer-Assisted and Robotic Endoscopy: Second International Workshop, CARE 2015, Held in Conjunction with MICCAI 2015, Munich, Germany, October 5, 2015, Revised Selected Papers. ed. / Xiongbiao Luo; Tobias Reichl; Austin Reiter ; Gian-Luca Mariottini . Vol. 9515 Springer Verlag, 2016. p. 129-139 (Lecture notes in computer science ; Vol. 9515).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - Enhancing normal-abnormal classification accuracy in colonoscopy videos via temporal consistency

AU - Puerto-Souza, Gustavo A.

AU - Manivannan, Siyamalan

AU - Trujillo, María P.

AU - Hoyos, Jesus A.

AU - Trucco, Emanuele

AU - Mariottini, Gian-Luca

PY - 2016/2/20

Y1 - 2016/2/20

N2 - This paper proposes a novel hierarchical approach to improve the accuracy of the classification of normal-vs-abnormal frames in white-light colonoscopy videos. The existing approaches label each frame independently, without considering the temporal consistency between adjacent frames. Temporal consistency, however, can improve the classification accuracy in the presence of unclear/uncertain images. We propose to leverage temporal consistency between adjacent frames for colonoscopy video frame classification using a novel hierarchical classifier. Comparative experiments with five challenging full colonoscopy videos show that the proposed approach considerably improves the mean class normal/abnormal classification accuracy compared to the approaches where the frames are classified independently.

AB - This paper proposes a novel hierarchical approach to improve the accuracy of the classification of normal-vs-abnormal frames in white-light colonoscopy videos. The existing approaches label each frame independently, without considering the temporal consistency between adjacent frames. Temporal consistency, however, can improve the classification accuracy in the presence of unclear/uncertain images. We propose to leverage temporal consistency between adjacent frames for colonoscopy video frame classification using a novel hierarchical classifier. Comparative experiments with five challenging full colonoscopy videos show that the proposed approach considerably improves the mean class normal/abnormal classification accuracy compared to the approaches where the frames are classified independently.

UR - http://www.scopus.com/inward/record.url?scp=84961214834&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-29965-5_13

DO - 10.1007/978-3-319-29965-5_13

M3 - Conference contribution

AN - SCOPUS:84961214834

SN - 9783319299648

VL - 9515

T3 - Lecture notes in computer science

SP - 129

EP - 139

BT - Computer-Assisted and Robotic Endoscopy

A2 - Luo, Xiongbiao

A2 - Reichl, Tobias

A2 - Reiter , Austin

A2 - Mariottini , Gian-Luca

PB - Springer Verlag

T2 - 2nd International Workshop on Computer-Assisted and Robotic Endoscopy

Y2 - 5 October 2015 through 5 October 2015

ER -

Puerto-Souza GA, Manivannan S, Trujillo MP, Hoyos JA, Trucco E, Mariottini GL. Enhancing normal-abnormal classification accuracy in colonoscopy videos via temporal consistency. In Luo X, Reichl T, Reiter A, Mariottini GL, editors, Computer-Assisted and Robotic Endoscopy: Second International Workshop, CARE 2015, Held in Conjunction with MICCAI 2015, Munich, Germany, October 5, 2015, Revised Selected Papers. Vol. 9515. Springer Verlag. 2016. p. 129-139. (Lecture notes in computer science ). doi: 10.1007/978-3-319-29965-5_13

Enhancing normal-abnormal classification accuracy in colonoscopy videos via temporal consistency

Abstract

Publication series

Workshop

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this