eprintid: 10122492 rev_number: 20 eprint_status: archive userid: 608 dir: disk0/10/12/24/92 datestamp: 2021-02-26 17:42:59 lastmod: 2024-10-22 10:41:59 status_changed: 2021-02-26 17:42:59 type: article metadata_visibility: show creators_name: Brandao, P creators_name: Psychogyios, D creators_name: Mazomenos, E creators_name: Stoyanov, D creators_name: Janatka, M title: HAPNet: hierarchically aggregated pyramid network for real-time stereo matching ispublished: pub divisions: UCL divisions: B04 divisions: C05 divisions: F48 divisions: F42 keywords: Convolutional neural networks, colonoscopy, computer-aided diagnosis note: This version is the author accepted manuscript. For information on re-use, please refer to the publisher’s terms and conditions. abstract: Recovering the 3D shape of the surgical site is crucial for multiple computer-assisted interventions. Stereo endoscopes can be used to compute 3D depth but computational stereo is a challenging, non-convex and inherently discontinuous optimisation problem. In this paper, we propose a deep learning architecture which avoids the explicit construction of a cost volume of similarity which is one of the most computationally costly blocks of stereo algorithms. This makes training our network significantly more efficient and avoids the needs for large memory allocation. Our method performs well, especially around regions comprising multiple discontinuities around surgical instrumentation or around complex small structures and instruments. The method compares well to the state-of-the-art techniques while taking a different methodological angle to computational stereo problem in surgical video. date: 2021 date_type: published official_url: https://dx.doi.org/10.1080/21681163.2020.1835561 oa_status: green full_text_type: other language: eng primo: open primo_central: open_green verified: verified_manual elements_id: 1834030 doi: 10.1080/21681163.2020.1835561 lyricists_name: Janatka, Miroslav lyricists_name: Mazomenos, Evangelos lyricists_name: Psychogyios, Dimitrios lyricists_name: Stoyanov, Danail lyricists_id: MJANA18 lyricists_id: EMAZO45 lyricists_id: DPSYC66 lyricists_id: DSTOY26 actors_name: Mazomenos, Evangelos actors_id: EMAZO45 actors_role: owner full_text_status: public publication: Computer Methods in Biomechanics and Biomedical Engineering: Imaging and Visualization volume: 9 number: 3 pagerange: 219-224 issn: 2168-1171 citation: Brandao, P; Psychogyios, D; Mazomenos, E; Stoyanov, D; Janatka, M; (2021) HAPNet: hierarchically aggregated pyramid network for real-time stereo matching. Computer Methods in Biomechanics and Biomedical Engineering: Imaging and Visualization , 9 (3) pp. 219-224. 10.1080/21681163.2020.1835561 <https://doi.org/10.1080/21681163.2020.1835561>. Green open access document_url: https://discovery.ucl.ac.uk/id/eprint/10122492/1/Computer%20Methods%20in%20Biomechanics%20and%20Biomedical%20Engineering%20Imaging%20and%20Visualization_2020_accepted.pdf