Real-Time Instrument Scene Detection in Screening GI Endoscopic Procedures

Chuanhai Zhang; Wallapak Tavanapong; Johnny Wong; Piet C.De Groen; Junghwan Oh

doi:10.1109/CBMS.2017.42

Real-Time Instrument Scene Detection in Screening GI Endoscopic Procedures

Chuanhai Zhang, Wallapak Tavanapong, Johnny Wong, Piet C.De Groen, Junghwan Oh

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

1 Scopus citations

Abstract

We describe a new and effective real-time solution for detecting video segments showing an instrument used during diagnostic or therapeutic operations in endoscopic procedures. In addition, we present a new method to collect a large training dataset: similarity-based data augmentation. This method automates most of the creation of a large training dataset and prevents extensive manual effort to collect and annotate training data by domain experts. Convolutional Neural Network (CNN) analysis using the training data collected with similarity-based data augmentation yields an average F1 score within 1% of that of the CNN analysis using a large manually collected training dataset.

Original language	English (US)
Title of host publication	Proceedings - 2017 IEEE 30th International Symposium on Computer-Based Medical Systems, CBMS 2017
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	720-725
Number of pages	6
Volume	2017-June
ISBN (Electronic)	9781538617106
DOIs	https://doi.org/10.1109/CBMS.2017.42
State	Published - Nov 10 2017
Externally published	Yes
Event	30th IEEE International Symposium on Computer-Based Medical Systems, CBMS 2017 - Thessaloniki, Greece Duration: Jun 22 2017 → Jun 24 2017

Other

Other	30th IEEE International Symposium on Computer-Based Medical Systems, CBMS 2017
Country	Greece
City	Thessaloniki
Period	6/22/17 → 6/24/17

Keywords

CNN
Scene Detection
Similarity Learning

ASJC Scopus subject areas

Radiology Nuclear Medicine and imaging
Computer Science Applications

Access to Document

10.1109/CBMS.2017.42

Fingerprint Dive into the research topics of 'Real-Time Instrument Scene Detection in Screening GI Endoscopic Procedures'. Together they form a unique fingerprint.

Cite this

Zhang, C., Tavanapong, W., Wong, J., Groen, P. C. D., & Oh, J. (2017). Real-Time Instrument Scene Detection in Screening GI Endoscopic Procedures. In Proceedings - 2017 IEEE 30th International Symposium on Computer-Based Medical Systems, CBMS 2017 (Vol. 2017-June, pp. 720-725). [8104286] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/CBMS.2017.42

Real-Time Instrument Scene Detection in Screening GI Endoscopic Procedures. / Zhang, Chuanhai; Tavanapong, Wallapak; Wong, Johnny; Groen, Piet C.De; Oh, Junghwan.

Proceedings - 2017 IEEE 30th International Symposium on Computer-Based Medical Systems, CBMS 2017. Vol. 2017-June Institute of Electrical and Electronics Engineers Inc., 2017. p. 720-725 8104286.

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

Zhang, C, Tavanapong, W, Wong, J, Groen, PCD & Oh, J 2017, Real-Time Instrument Scene Detection in Screening GI Endoscopic Procedures. in Proceedings - 2017 IEEE 30th International Symposium on Computer-Based Medical Systems, CBMS 2017. vol. 2017-June, 8104286, Institute of Electrical and Electronics Engineers Inc., pp. 720-725, 30th IEEE International Symposium on Computer-Based Medical Systems, CBMS 2017, Thessaloniki, Greece, 6/22/17. https://doi.org/10.1109/CBMS.2017.42

@inproceedings{557d698282a74dbbbe8ab5d6e73202b8,

title = "Real-Time Instrument Scene Detection in Screening GI Endoscopic Procedures",

abstract = "We describe a new and effective real-time solution for detecting video segments showing an instrument used during diagnostic or therapeutic operations in endoscopic procedures. In addition, we present a new method to collect a large training dataset: similarity-based data augmentation. This method automates most of the creation of a large training dataset and prevents extensive manual effort to collect and annotate training data by domain experts. Convolutional Neural Network (CNN) analysis using the training data collected with similarity-based data augmentation yields an average F1 score within 1% of that of the CNN analysis using a large manually collected training dataset.",

keywords = "CNN, Scene Detection, Similarity Learning",

author = "Chuanhai Zhang and Wallapak Tavanapong and Johnny Wong and Groen, {Piet C.De} and Junghwan Oh",

year = "2017",

month = nov,

day = "10",

doi = "10.1109/CBMS.2017.42",

language = "English (US)",

volume = "2017-June",

pages = "720--725",

booktitle = "Proceedings - 2017 IEEE 30th International Symposium on Computer-Based Medical Systems, CBMS 2017",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

note = "30th IEEE International Symposium on Computer-Based Medical Systems, CBMS 2017 ; Conference date: 22-06-2017 Through 24-06-2017",

}

TY - GEN

T1 - Real-Time Instrument Scene Detection in Screening GI Endoscopic Procedures

AU - Zhang, Chuanhai

AU - Tavanapong, Wallapak

AU - Wong, Johnny

AU - Groen, Piet C.De

AU - Oh, Junghwan

PY - 2017/11/10

Y1 - 2017/11/10

N2 - We describe a new and effective real-time solution for detecting video segments showing an instrument used during diagnostic or therapeutic operations in endoscopic procedures. In addition, we present a new method to collect a large training dataset: similarity-based data augmentation. This method automates most of the creation of a large training dataset and prevents extensive manual effort to collect and annotate training data by domain experts. Convolutional Neural Network (CNN) analysis using the training data collected with similarity-based data augmentation yields an average F1 score within 1% of that of the CNN analysis using a large manually collected training dataset.

AB - We describe a new and effective real-time solution for detecting video segments showing an instrument used during diagnostic or therapeutic operations in endoscopic procedures. In addition, we present a new method to collect a large training dataset: similarity-based data augmentation. This method automates most of the creation of a large training dataset and prevents extensive manual effort to collect and annotate training data by domain experts. Convolutional Neural Network (CNN) analysis using the training data collected with similarity-based data augmentation yields an average F1 score within 1% of that of the CNN analysis using a large manually collected training dataset.

KW - CNN

KW - Scene Detection

KW - Similarity Learning

UR - http://www.scopus.com/inward/record.url?scp=85040342841&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85040342841&partnerID=8YFLogxK

U2 - 10.1109/CBMS.2017.42

DO - 10.1109/CBMS.2017.42

M3 - Conference contribution

AN - SCOPUS:85040342841

VL - 2017-June

SP - 720

EP - 725

BT - Proceedings - 2017 IEEE 30th International Symposium on Computer-Based Medical Systems, CBMS 2017

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 30th IEEE International Symposium on Computer-Based Medical Systems, CBMS 2017

Y2 - 22 June 2017 through 24 June 2017

ER -