Uninformative Frame Detection in Colonoscopy Through Motion, Edge and Color Features

Ali ARMIN, Girija CHETTY, Jurgen Fripp, Hans De Visser, Cedric Dumas, Amir Fazlollahi, Florian Grimpen, olivier salvado

Research output: A Conference proceeding or a Chapter in BookConference contribution

4 Citations (Scopus)

Abstract

Colonoscopy is performed by using a long endoscope inserted in the colon of patients to inspect the internal mucosa. During the intervention, clinicians observe the colon under bright light to diagnose pathology and guide intervention. We are developing a computer aided system to facilitate navigation and diagnosis. One essential step is to estimate the camera pose relative to the colon from video frames. However, within every colonoscopy video is a large number of frames that provide no structural information (e.g. blurry or out of focus frames or those close to the colon wall). This hampers our camera pose estimation algorithm. To distinguish uninformative frames from informative ones, we investigated several features computed from each frame: corner and edge features matched with the previous frame, the percentage of edge pixels, and the mean and standard deviation of intensity in hue-saturation-value color space. A Random Forest classifier was used for classification. The method was validated on four colonoscopy videos that were manually classified. The resulting classification had a sensitivity of 75 % and specificity of 97 % for detecting uninformative frames. The proposed features not only compared favorably to existing techniques for detecting uninformative frames, but they also can be utilized for the camera navigation purpose
Original languageEnglish
Title of host publicationComputer-Assisted and Robotic Endoscopy
Subtitle of host publicationSecond International Workshop, CARE 2015
EditorsXiongbiao Luo, Tobias Reichl, Austin Reiter, Gian-Luca Mariottini
Place of PublicationMunich, Germany
PublisherSpringer
Pages153-162
Number of pages10
Volume9515
ISBN (Electronic)9783319299655
ISBN (Print)9783319299648
DOIs
Publication statusPublished - 2016
EventCARE 2015, 2nd International Workshop on Computer Assisted and Robotic Endoscopy - Munich, Munich, Germany
Duration: 5 Oct 20159 Oct 2015

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume9515
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

ConferenceCARE 2015, 2nd International Workshop on Computer Assisted and Robotic Endoscopy
Abbreviated titleCARE 2015
CountryGermany
CityMunich
Period5/10/159/10/15

Fingerprint

Cameras
Color
Motion
Navigation
Endoscopy
Pathology
Camera
Classifiers
Pixels
Endoscope
Mean deviation
Pose Estimation
Random Forest
Color Space
Estimation Algorithms
Standard deviation
Specificity
Percentage
Saturation
Pixel

Cite this

ARMIN, A., CHETTY, G., Fripp, J., De Visser, H., Dumas, C., Fazlollahi, A., ... salvado, O. (2016). Uninformative Frame Detection in Colonoscopy Through Motion, Edge and Color Features. In X. Luo, T. Reichl, A. Reiter, & G-L. Mariottini (Eds.), Computer-Assisted and Robotic Endoscopy: Second International Workshop, CARE 2015 (Vol. 9515, pp. 153-162). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9515). Munich, Germany: Springer. https://doi.org/10.1007/978-3-319-29965-5_15
ARMIN, Ali ; CHETTY, Girija ; Fripp, Jurgen ; De Visser, Hans ; Dumas, Cedric ; Fazlollahi, Amir ; Grimpen, Florian ; salvado, olivier. / Uninformative Frame Detection in Colonoscopy Through Motion, Edge and Color Features. Computer-Assisted and Robotic Endoscopy: Second International Workshop, CARE 2015. editor / Xiongbiao Luo ; Tobias Reichl ; Austin Reiter ; Gian-Luca Mariottini. Vol. 9515 Munich, Germany : Springer, 2016. pp. 153-162 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{2c7802749fdb42309844f77e3ab4f20e,
title = "Uninformative Frame Detection in Colonoscopy Through Motion, Edge and Color Features",
abstract = "Colonoscopy is performed by using a long endoscope inserted in the colon of patients to inspect the internal mucosa. During the intervention, clinicians observe the colon under bright light to diagnose pathology and guide intervention. We are developing a computer aided system to facilitate navigation and diagnosis. One essential step is to estimate the camera pose relative to the colon from video frames. However, within every colonoscopy video is a large number of frames that provide no structural information (e.g. blurry or out of focus frames or those close to the colon wall). This hampers our camera pose estimation algorithm. To distinguish uninformative frames from informative ones, we investigated several features computed from each frame: corner and edge features matched with the previous frame, the percentage of edge pixels, and the mean and standard deviation of intensity in hue-saturation-value color space. A Random Forest classifier was used for classification. The method was validated on four colonoscopy videos that were manually classified. The resulting classification had a sensitivity of 75 {\%} and specificity of 97 {\%} for detecting uninformative frames. The proposed features not only compared favorably to existing techniques for detecting uninformative frames, but they also can be utilized for the camera navigation purpose",
keywords = "optical-colonoscopy, colonoscopy-quality, uninformative-frames, Uninformative frames, Optical colonoscopy, Colonoscopy quality, Feature, Random Forest",
author = "Ali ARMIN and Girija CHETTY and Jurgen Fripp and {De Visser}, Hans and Cedric Dumas and Amir Fazlollahi and Florian Grimpen and olivier salvado",
year = "2016",
doi = "10.1007/978-3-319-29965-5_15",
language = "English",
isbn = "9783319299648",
volume = "9515",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer",
pages = "153--162",
editor = "Xiongbiao Luo and Tobias Reichl and Austin Reiter and Gian-Luca Mariottini",
booktitle = "Computer-Assisted and Robotic Endoscopy",
address = "Netherlands",

}

ARMIN, A, CHETTY, G, Fripp, J, De Visser, H, Dumas, C, Fazlollahi, A, Grimpen, F & salvado, O 2016, Uninformative Frame Detection in Colonoscopy Through Motion, Edge and Color Features. in X Luo, T Reichl, A Reiter & G-L Mariottini (eds), Computer-Assisted and Robotic Endoscopy: Second International Workshop, CARE 2015. vol. 9515, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 9515, Springer, Munich, Germany, pp. 153-162, CARE 2015, 2nd International Workshop on Computer Assisted and Robotic Endoscopy, Munich, Germany, 5/10/15. https://doi.org/10.1007/978-3-319-29965-5_15

Uninformative Frame Detection in Colonoscopy Through Motion, Edge and Color Features. / ARMIN, Ali; CHETTY, Girija; Fripp, Jurgen; De Visser, Hans; Dumas, Cedric; Fazlollahi, Amir; Grimpen, Florian; salvado, olivier.

Computer-Assisted and Robotic Endoscopy: Second International Workshop, CARE 2015. ed. / Xiongbiao Luo; Tobias Reichl; Austin Reiter; Gian-Luca Mariottini. Vol. 9515 Munich, Germany : Springer, 2016. p. 153-162 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9515).

Research output: A Conference proceeding or a Chapter in BookConference contribution

TY - GEN

T1 - Uninformative Frame Detection in Colonoscopy Through Motion, Edge and Color Features

AU - ARMIN, Ali

AU - CHETTY, Girija

AU - Fripp, Jurgen

AU - De Visser, Hans

AU - Dumas, Cedric

AU - Fazlollahi, Amir

AU - Grimpen, Florian

AU - salvado, olivier

PY - 2016

Y1 - 2016

N2 - Colonoscopy is performed by using a long endoscope inserted in the colon of patients to inspect the internal mucosa. During the intervention, clinicians observe the colon under bright light to diagnose pathology and guide intervention. We are developing a computer aided system to facilitate navigation and diagnosis. One essential step is to estimate the camera pose relative to the colon from video frames. However, within every colonoscopy video is a large number of frames that provide no structural information (e.g. blurry or out of focus frames or those close to the colon wall). This hampers our camera pose estimation algorithm. To distinguish uninformative frames from informative ones, we investigated several features computed from each frame: corner and edge features matched with the previous frame, the percentage of edge pixels, and the mean and standard deviation of intensity in hue-saturation-value color space. A Random Forest classifier was used for classification. The method was validated on four colonoscopy videos that were manually classified. The resulting classification had a sensitivity of 75 % and specificity of 97 % for detecting uninformative frames. The proposed features not only compared favorably to existing techniques for detecting uninformative frames, but they also can be utilized for the camera navigation purpose

AB - Colonoscopy is performed by using a long endoscope inserted in the colon of patients to inspect the internal mucosa. During the intervention, clinicians observe the colon under bright light to diagnose pathology and guide intervention. We are developing a computer aided system to facilitate navigation and diagnosis. One essential step is to estimate the camera pose relative to the colon from video frames. However, within every colonoscopy video is a large number of frames that provide no structural information (e.g. blurry or out of focus frames or those close to the colon wall). This hampers our camera pose estimation algorithm. To distinguish uninformative frames from informative ones, we investigated several features computed from each frame: corner and edge features matched with the previous frame, the percentage of edge pixels, and the mean and standard deviation of intensity in hue-saturation-value color space. A Random Forest classifier was used for classification. The method was validated on four colonoscopy videos that were manually classified. The resulting classification had a sensitivity of 75 % and specificity of 97 % for detecting uninformative frames. The proposed features not only compared favorably to existing techniques for detecting uninformative frames, but they also can be utilized for the camera navigation purpose

KW - optical-colonoscopy

KW - colonoscopy-quality

KW - uninformative-frames

KW - Uninformative frames

KW - Optical colonoscopy

KW - Colonoscopy quality

KW - Feature

KW - Random Forest

UR - http://www.scopus.com/inward/record.url?scp=84961223475&partnerID=8YFLogxK

UR - http://www.mendeley.com/research/uninformative-frame-detection-colonoscopy-through-motion-edge-color-features

U2 - 10.1007/978-3-319-29965-5_15

DO - 10.1007/978-3-319-29965-5_15

M3 - Conference contribution

SN - 9783319299648

VL - 9515

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 153

EP - 162

BT - Computer-Assisted and Robotic Endoscopy

A2 - Luo, Xiongbiao

A2 - Reichl, Tobias

A2 - Reiter, Austin

A2 - Mariottini, Gian-Luca

PB - Springer

CY - Munich, Germany

ER -

ARMIN A, CHETTY G, Fripp J, De Visser H, Dumas C, Fazlollahi A et al. Uninformative Frame Detection in Colonoscopy Through Motion, Edge and Color Features. In Luo X, Reichl T, Reiter A, Mariottini G-L, editors, Computer-Assisted and Robotic Endoscopy: Second International Workshop, CARE 2015. Vol. 9515. Munich, Germany: Springer. 2016. p. 153-162. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-29965-5_15