Curriculum deep reinforcement learning with different exploration strategies : a feasibility study on cardiac landmark detection
- Author
- Patricio Astudillo, Peter Mortier, Matthieu De Beule and Francis wyffels (UGent)
- Organization
- Abstract
- Transcatheter aortic valve implantation (TAVI) is associated with conduction abnormalities and the mechanical interaction between the prosthesis and the atrioventricular (AV) conduction path cause these life-threatening arrhythmias. Pre-operative assessment of the location of the AV conduction path can help to understand the risk of post-TAVI conduction abnormalities. As the AV conduction path is not visible on cardiac CT, the inferior border of the membranous septum can be used as an anatomical landmark. Detecting this border automatically, accurately and efficiently would save operator time and thus benefit pre-operative planning. This preliminary study was performed to identify the feasibility of 3D landmark detection in cardiac CT images with curriculum deep Q-learning. In this study, curriculum learning was used to gradually teach an artificial agent to detect this anatomical landmark from cardiac CT. This agent was equipped with a small field of view and burdened with a large ac tion-space. Moreover, we introduced two novel action-selection strategies: α-decay and action-dropout. We compared these two strategies to the already established ε-decay strategy and observed that α-decay yielded the most accurate results. Limited computational resources were used to ensure reproducibility. In order to maximize the amount of patient data, the method was cross-validated with k-folding for all three action-selection strategies. An inter-operator variability study was conducted to assess the accuracy of the method
- Keywords
- Biomedical Informatics, Cardiography, Medical Information Systems, Semi-supervised Learning, AORTIC-VALVE-REPLACEMENT, TRANSCATHETER, RISK
Downloads
-
published 8637667.pdf
- full text (Accepted manuscript)
- |
- open access
- |
- |
- 605.89 KB
Citation
Please use this url to cite or link to this publication: http://hdl.handle.net/1854/LU-8637667
- MLA
- Astudillo, Patricio, et al. “Curriculum Deep Reinforcement Learning with Different Exploration Strategies : A Feasibility Study on Cardiac Landmark Detection.” PROCEEDINGS OF THE 13TH INTERNATIONAL JOINT CONFERENCE ON BIOMEDICAL ENGINEERING SYSTEMS AND TECHNOLOGIES, VOL 2: BIOIMAGING, edited by Filipe Soares et al., Scitepress, 2020, pp. 37–45, doi:10.5220/0008948900370045.
- APA
- Astudillo, P., Mortier, P., De Beule, M., & wyffels, F. (2020). Curriculum deep reinforcement learning with different exploration strategies : a feasibility study on cardiac landmark detection. In F. Soares, A. Fred, & H. Gamboa (Eds.), PROCEEDINGS OF THE 13TH INTERNATIONAL JOINT CONFERENCE ON BIOMEDICAL ENGINEERING SYSTEMS AND TECHNOLOGIES, VOL 2: BIOIMAGING (pp. 37–45). https://doi.org/10.5220/0008948900370045
- Chicago author-date
- Astudillo, Patricio, Peter Mortier, Matthieu De Beule, and Francis wyffels. 2020. “Curriculum Deep Reinforcement Learning with Different Exploration Strategies : A Feasibility Study on Cardiac Landmark Detection.” In PROCEEDINGS OF THE 13TH INTERNATIONAL JOINT CONFERENCE ON BIOMEDICAL ENGINEERING SYSTEMS AND TECHNOLOGIES, VOL 2: BIOIMAGING, edited by Filipe Soares, Ana Fred, and Hugo Gamboa, 37–45. Scitepress. https://doi.org/10.5220/0008948900370045.
- Chicago author-date (all authors)
- Astudillo, Patricio, Peter Mortier, Matthieu De Beule, and Francis wyffels. 2020. “Curriculum Deep Reinforcement Learning with Different Exploration Strategies : A Feasibility Study on Cardiac Landmark Detection.” In PROCEEDINGS OF THE 13TH INTERNATIONAL JOINT CONFERENCE ON BIOMEDICAL ENGINEERING SYSTEMS AND TECHNOLOGIES, VOL 2: BIOIMAGING, ed by. Filipe Soares, Ana Fred, and Hugo Gamboa, 37–45. Scitepress. doi:10.5220/0008948900370045.
- Vancouver
- 1.Astudillo P, Mortier P, De Beule M, wyffels F. Curriculum deep reinforcement learning with different exploration strategies : a feasibility study on cardiac landmark detection. In: Soares F, Fred A, Gamboa H, editors. PROCEEDINGS OF THE 13TH INTERNATIONAL JOINT CONFERENCE ON BIOMEDICAL ENGINEERING SYSTEMS AND TECHNOLOGIES, VOL 2: BIOIMAGING. Scitepress; 2020. p. 37–45.
- IEEE
- [1]P. Astudillo, P. Mortier, M. De Beule, and F. wyffels, “Curriculum deep reinforcement learning with different exploration strategies : a feasibility study on cardiac landmark detection,” in PROCEEDINGS OF THE 13TH INTERNATIONAL JOINT CONFERENCE ON BIOMEDICAL ENGINEERING SYSTEMS AND TECHNOLOGIES, VOL 2: BIOIMAGING, Valetta, Malta, 2020, pp. 37–45.
@inproceedings{8637667,
abstract = {{Transcatheter aortic valve implantation (TAVI) is associated with conduction abnormalities and the mechanical interaction between the prosthesis and the atrioventricular (AV) conduction path cause these life-threatening arrhythmias. Pre-operative assessment of the location of the AV conduction path can help to understand the risk of post-TAVI conduction abnormalities. As the AV conduction path is not visible on cardiac CT, the inferior border of the membranous septum can be used as an anatomical landmark. Detecting this border automatically, accurately and efficiently would save operator time and thus benefit pre-operative planning. This preliminary study was performed to identify the feasibility of 3D landmark detection in cardiac CT images with curriculum deep Q-learning. In this study, curriculum learning was used to gradually teach an artificial agent to detect this anatomical landmark from cardiac CT. This agent was equipped with a small field of view and burdened with a large ac tion-space. Moreover, we introduced two novel action-selection strategies: α-decay and action-dropout. We compared these two strategies to the already established ε-decay strategy and observed that α-decay yielded the most accurate results. Limited computational resources were used to ensure reproducibility. In order to maximize the amount of patient data, the method was cross-validated with k-folding for all three action-selection strategies. An inter-operator variability study was conducted to assess the accuracy of the method}},
author = {{Astudillo, Patricio and Mortier, Peter and De Beule, Matthieu and wyffels, Francis}},
booktitle = {{PROCEEDINGS OF THE 13TH INTERNATIONAL JOINT CONFERENCE ON BIOMEDICAL ENGINEERING SYSTEMS AND TECHNOLOGIES, VOL 2: BIOIMAGING}},
editor = {{Soares, Filipe and Fred, Ana and Gamboa, Hugo}},
isbn = {{9789897583988}},
keywords = {{Biomedical Informatics,Cardiography,Medical Information Systems,Semi-supervised Learning,AORTIC-VALVE-REPLACEMENT,TRANSCATHETER,RISK}},
language = {{eng}},
location = {{Valetta, Malta}},
pages = {{37--45}},
publisher = {{Scitepress}},
title = {{Curriculum deep reinforcement learning with different exploration strategies : a feasibility study on cardiac landmark detection}},
url = {{http://doi.org/10.5220/0008948900370045}},
year = {{2020}},
}
- Altmetric
- View in Altmetric
- Web of Science
- Times cited: