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