Guided Unsupervised Desmoking of Laparoscopic Images Using Cycle-Desmoke

V. Vishal, Neeraj Sharma, Munendra Singh

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

The generation of smoke in laparoscopic surgery due to laser ablation and cauterization causes deterioration in the visual quality of the operative field. In order to reduce the effect of smoke, the present paper proposes an end-to-end network, called Cycle-Desmoke. The network enhances the CycleGAN framework by adoption of a new generator architecture and addition of new Guided-Unsharp Upsample loss in combination to adversarial and cycle-consistency loss. The Atrous Convolution Feature Extraction Module present in the encoder blocks of the generator helps distinguishing smoke by capturing features at multiple scales by the use of kernels with different receptive fields. Further, the use of Guided-Unsharp Upsample loss supervises the upsampling process of the feature maps and helps improve the contrast of the desmoked image. The network performs robust unsupervised Image-to-Image Translation from smoke domain to smoke-free domain. The public Cholec80 dataset is used to evaluate the performance of the proposed method. Quantitative and qualitative comparative analysis of the proposed method over the state-of-the-methods reveals the effectiveness of the method at the task of smoke removal and enhancement of the image.

Original languageEnglish
Title of host publicationOR 2.0 Context-Aware Operating Theaters and Machine Learning in Clinical Neuroimaging - 2nd International Workshop, OR 2.0 2019, and 2nd International Workshop, MLCN 2019, Held in Conjunction with MICCAI 2019, Proceedings
EditorsLuping Zhou, Duygu Sarikaya, Seyed Mostafa Kia, Stefanie Speidel, Anand Malpani, Daniel Hashimoto, Mohamad Habes, Tommy Löfstedt, Kerstin Ritter, Hongzhi Wang
PublisherSpringer Paris
Pages21-28
Number of pages8
ISBN (Print)9783030326944
DOIs
Publication statusPublished - 01-01-2019
Externally publishedYes
Event2nd International Workshop on Context-Aware Surgical Theaters, OR 2.0 2019, and the 2nd International Workshop on Machine Learning in Clinical Neuroimaging, MLCN 2019, held in conjunction with the 22nd International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2019 - Shenzhen, China
Duration: 17-10-201917-10-2019

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11796 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference2nd International Workshop on Context-Aware Surgical Theaters, OR 2.0 2019, and the 2nd International Workshop on Machine Learning in Clinical Neuroimaging, MLCN 2019, held in conjunction with the 22nd International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2019
CountryChina
CityShenzhen
Period17-10-1917-10-19

Fingerprint

Smoke
Cycle
Generator
Receptive Field
Laser Ablation
Multiple Scales
Qualitative Analysis
Encoder
Deterioration
Comparative Analysis
Surgery
Feature Extraction
Convolution
Enhancement
Laser ablation
kernel
Module
Feature extraction
Evaluate

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Vishal, V., Sharma, N., & Singh, M. (2019). Guided Unsupervised Desmoking of Laparoscopic Images Using Cycle-Desmoke. In L. Zhou, D. Sarikaya, S. M. Kia, S. Speidel, A. Malpani, D. Hashimoto, M. Habes, T. Löfstedt, K. Ritter, ... H. Wang (Eds.), OR 2.0 Context-Aware Operating Theaters and Machine Learning in Clinical Neuroimaging - 2nd International Workshop, OR 2.0 2019, and 2nd International Workshop, MLCN 2019, Held in Conjunction with MICCAI 2019, Proceedings (pp. 21-28). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11796 LNCS). Springer Paris. https://doi.org/10.1007/978-3-030-32695-1_3
Vishal, V. ; Sharma, Neeraj ; Singh, Munendra. / Guided Unsupervised Desmoking of Laparoscopic Images Using Cycle-Desmoke. OR 2.0 Context-Aware Operating Theaters and Machine Learning in Clinical Neuroimaging - 2nd International Workshop, OR 2.0 2019, and 2nd International Workshop, MLCN 2019, Held in Conjunction with MICCAI 2019, Proceedings. editor / Luping Zhou ; Duygu Sarikaya ; Seyed Mostafa Kia ; Stefanie Speidel ; Anand Malpani ; Daniel Hashimoto ; Mohamad Habes ; Tommy Löfstedt ; Kerstin Ritter ; Hongzhi Wang. Springer Paris, 2019. pp. 21-28 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{d2c3c15461d94b8db1bb2d8376ce5f74,
title = "Guided Unsupervised Desmoking of Laparoscopic Images Using Cycle-Desmoke",
abstract = "The generation of smoke in laparoscopic surgery due to laser ablation and cauterization causes deterioration in the visual quality of the operative field. In order to reduce the effect of smoke, the present paper proposes an end-to-end network, called Cycle-Desmoke. The network enhances the CycleGAN framework by adoption of a new generator architecture and addition of new Guided-Unsharp Upsample loss in combination to adversarial and cycle-consistency loss. The Atrous Convolution Feature Extraction Module present in the encoder blocks of the generator helps distinguishing smoke by capturing features at multiple scales by the use of kernels with different receptive fields. Further, the use of Guided-Unsharp Upsample loss supervises the upsampling process of the feature maps and helps improve the contrast of the desmoked image. The network performs robust unsupervised Image-to-Image Translation from smoke domain to smoke-free domain. The public Cholec80 dataset is used to evaluate the performance of the proposed method. Quantitative and qualitative comparative analysis of the proposed method over the state-of-the-methods reveals the effectiveness of the method at the task of smoke removal and enhancement of the image.",
author = "V. Vishal and Neeraj Sharma and Munendra Singh",
year = "2019",
month = "1",
day = "1",
doi = "10.1007/978-3-030-32695-1_3",
language = "English",
isbn = "9783030326944",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Paris",
pages = "21--28",
editor = "Luping Zhou and Duygu Sarikaya and Kia, {Seyed Mostafa} and Stefanie Speidel and Anand Malpani and Daniel Hashimoto and Mohamad Habes and Tommy L{\"o}fstedt and Kerstin Ritter and Hongzhi Wang",
booktitle = "OR 2.0 Context-Aware Operating Theaters and Machine Learning in Clinical Neuroimaging - 2nd International Workshop, OR 2.0 2019, and 2nd International Workshop, MLCN 2019, Held in Conjunction with MICCAI 2019, Proceedings",
address = "France",

}

Vishal, V, Sharma, N & Singh, M 2019, Guided Unsupervised Desmoking of Laparoscopic Images Using Cycle-Desmoke. in L Zhou, D Sarikaya, SM Kia, S Speidel, A Malpani, D Hashimoto, M Habes, T Löfstedt, K Ritter & H Wang (eds), OR 2.0 Context-Aware Operating Theaters and Machine Learning in Clinical Neuroimaging - 2nd International Workshop, OR 2.0 2019, and 2nd International Workshop, MLCN 2019, Held in Conjunction with MICCAI 2019, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11796 LNCS, Springer Paris, pp. 21-28, 2nd International Workshop on Context-Aware Surgical Theaters, OR 2.0 2019, and the 2nd International Workshop on Machine Learning in Clinical Neuroimaging, MLCN 2019, held in conjunction with the 22nd International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2019, Shenzhen, China, 17-10-19. https://doi.org/10.1007/978-3-030-32695-1_3

Guided Unsupervised Desmoking of Laparoscopic Images Using Cycle-Desmoke. / Vishal, V.; Sharma, Neeraj; Singh, Munendra.

OR 2.0 Context-Aware Operating Theaters and Machine Learning in Clinical Neuroimaging - 2nd International Workshop, OR 2.0 2019, and 2nd International Workshop, MLCN 2019, Held in Conjunction with MICCAI 2019, Proceedings. ed. / Luping Zhou; Duygu Sarikaya; Seyed Mostafa Kia; Stefanie Speidel; Anand Malpani; Daniel Hashimoto; Mohamad Habes; Tommy Löfstedt; Kerstin Ritter; Hongzhi Wang. Springer Paris, 2019. p. 21-28 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11796 LNCS).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - Guided Unsupervised Desmoking of Laparoscopic Images Using Cycle-Desmoke

AU - Vishal, V.

AU - Sharma, Neeraj

AU - Singh, Munendra

PY - 2019/1/1

Y1 - 2019/1/1

N2 - The generation of smoke in laparoscopic surgery due to laser ablation and cauterization causes deterioration in the visual quality of the operative field. In order to reduce the effect of smoke, the present paper proposes an end-to-end network, called Cycle-Desmoke. The network enhances the CycleGAN framework by adoption of a new generator architecture and addition of new Guided-Unsharp Upsample loss in combination to adversarial and cycle-consistency loss. The Atrous Convolution Feature Extraction Module present in the encoder blocks of the generator helps distinguishing smoke by capturing features at multiple scales by the use of kernels with different receptive fields. Further, the use of Guided-Unsharp Upsample loss supervises the upsampling process of the feature maps and helps improve the contrast of the desmoked image. The network performs robust unsupervised Image-to-Image Translation from smoke domain to smoke-free domain. The public Cholec80 dataset is used to evaluate the performance of the proposed method. Quantitative and qualitative comparative analysis of the proposed method over the state-of-the-methods reveals the effectiveness of the method at the task of smoke removal and enhancement of the image.

AB - The generation of smoke in laparoscopic surgery due to laser ablation and cauterization causes deterioration in the visual quality of the operative field. In order to reduce the effect of smoke, the present paper proposes an end-to-end network, called Cycle-Desmoke. The network enhances the CycleGAN framework by adoption of a new generator architecture and addition of new Guided-Unsharp Upsample loss in combination to adversarial and cycle-consistency loss. The Atrous Convolution Feature Extraction Module present in the encoder blocks of the generator helps distinguishing smoke by capturing features at multiple scales by the use of kernels with different receptive fields. Further, the use of Guided-Unsharp Upsample loss supervises the upsampling process of the feature maps and helps improve the contrast of the desmoked image. The network performs robust unsupervised Image-to-Image Translation from smoke domain to smoke-free domain. The public Cholec80 dataset is used to evaluate the performance of the proposed method. Quantitative and qualitative comparative analysis of the proposed method over the state-of-the-methods reveals the effectiveness of the method at the task of smoke removal and enhancement of the image.

UR - http://www.scopus.com/inward/record.url?scp=85075569490&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85075569490&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-32695-1_3

DO - 10.1007/978-3-030-32695-1_3

M3 - Conference contribution

AN - SCOPUS:85075569490

SN - 9783030326944

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

SP - 21

EP - 28

BT - OR 2.0 Context-Aware Operating Theaters and Machine Learning in Clinical Neuroimaging - 2nd International Workshop, OR 2.0 2019, and 2nd International Workshop, MLCN 2019, Held in Conjunction with MICCAI 2019, Proceedings

A2 - Zhou, Luping

A2 - Sarikaya, Duygu

A2 - Kia, Seyed Mostafa

A2 - Speidel, Stefanie

A2 - Malpani, Anand

A2 - Hashimoto, Daniel

A2 - Habes, Mohamad

A2 - Löfstedt, Tommy

A2 - Ritter, Kerstin

A2 - Wang, Hongzhi

PB - Springer Paris

ER -

Vishal V, Sharma N, Singh M. Guided Unsupervised Desmoking of Laparoscopic Images Using Cycle-Desmoke. In Zhou L, Sarikaya D, Kia SM, Speidel S, Malpani A, Hashimoto D, Habes M, Löfstedt T, Ritter K, Wang H, editors, OR 2.0 Context-Aware Operating Theaters and Machine Learning in Clinical Neuroimaging - 2nd International Workshop, OR 2.0 2019, and 2nd International Workshop, MLCN 2019, Held in Conjunction with MICCAI 2019, Proceedings. Springer Paris. 2019. p. 21-28. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-030-32695-1_3