eprintid: 10183589
rev_number: 6
eprint_status: archive
userid: 699
dir: disk0/10/18/35/89
datestamp: 2023-12-12 15:33:47
lastmod: 2023-12-12 15:33:47
status_changed: 2023-12-12 15:33:47
type: proceedings_section
metadata_visibility: show
sword_depositor: 699
creators_name: Yi, W
creators_name: Stavrinides, V
creators_name: Baum, ZMC
creators_name: Yang, Q
creators_name: Barratt, DC
creators_name: Clarkson, MJ
creators_name: Hu, Y
creators_name: Saeed, SU
title: Boundary-RL: Reinforcement Learning for Weakly-Supervised Prostate Segmentation in TRUS Images
ispublished: pub
divisions: UCL
divisions: B02
divisions: B04
divisions: C10
divisions: C05
divisions: D19
divisions: F42
divisions: G99
note: This version is the author accepted manuscript. For information on re-use, please refer to the publisher’s terms and conditions.
abstract: We propose Boundary-RL, a novel weakly supervised segmentation method that utilises only patch-level labels for training. We envision segmentation as a boundary detection problem, rather than a pixel-level classification as in previous works. This outlook on segmentation may allow for boundary delineation under challenging scenarios such as where noise artefacts may be present within the region-of-interest (ROI) boundaries, where traditional pixel-level classification-based weakly supervised methods may not be able to effectively segment the ROI. Particularly of interest, ultrasound images, where intensity values represent acoustic impedance differences between boundaries, may also benefit from the boundary delineation approach. Our method uses reinforcement learning to train a controller function to localise boundaries of ROIs using a reward derived from a pre-trained boundary-presence classifier. The classifier indicates when an object boundary is encountered within a patch, serving as weak supervision, as the controller modifies the patch location in a sequential Markov decision process. The classifier itself is trained using only binary patch-level labels of object presence, the only labels used during training of the entire boundary delineation framework. The use of a controller ensures that sliding window over the entire image is not necessary and reduces possible false-positives or -negatives by minimising number of patches passed to the boundary-presence classifier. We evaluate our approach for a clinically relevant task of prostate gland segmentation on trans-rectal ultrasound images. We show improved performance compared to other tested weakly supervised methods, using the same labels e.g., multiple instance learning.
date: 2023-10-05
date_type: published
publisher: Springer Nature
official_url: https://doi.org/10.1007/978-3-031-45673-2
oa_status: green
full_text_type: other
language: eng
primo: open
primo_central: open_green
verified: verified_manual
elements_id: 2108489
doi: 10.1007/978-3-031-45673-2_28
lyricists_name: Saeed, Shaheer Ullah
lyricists_name: Barratt, Dean
lyricists_name: Stavrinides, Vasilis
lyricists_name: Yang, Qianye
lyricists_name: Clarkson, Matthew
lyricists_id: SUSAE80
lyricists_id: DBARR55
lyricists_id: VSTAV68
lyricists_id: QYANA79
lyricists_id: MJCLA42
actors_name: Saeed, Shaheer Ullah
actors_id: SUSAE80
actors_role: owner
full_text_status: public
pres_type: paper
series: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
publication: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
volume: 14348
place_of_pub: Cham, Switzerland
pagerange: 277-288
event_title: MLMI 2023
book_title: Machine Learning in Medical Imaging
editors_name: Cao, Xiaohuan
editors_name: Xu, Xuanang
editors_name: Rekik, Islem
editors_name: Cui, Zhiming
editors_name: Ouyang, Xi
citation:        Yi, W;    Stavrinides, V;    Baum, ZMC;    Yang, Q;    Barratt, DC;    Clarkson, MJ;    Hu, Y;           Yi, W;  Stavrinides, V;  Baum, ZMC;  Yang, Q;  Barratt, DC;  Clarkson, MJ;  Hu, Y;  Saeed, SU;   - view fewer <#>    (2023)    Boundary-RL: Reinforcement Learning for Weakly-Supervised Prostate Segmentation in TRUS Images.                     In: Cao, Xiaohuan and Xu, Xuanang and Rekik, Islem and Cui, Zhiming and Ouyang, Xi, (eds.) Machine Learning in Medical Imaging.  (pp. pp. 277-288).  Springer Nature: Cham, Switzerland.       Green open access   
 
document_url: https://discovery.ucl.ac.uk/id/eprint/10183589/1/MLMI2023_Boundary_RL-3.pdf