Přístupnostní navigace
E-přihláška
Vyhledávání Vyhledat Zavřít
Detail publikace
VIČAR, T. CHMELÍK, J. JAKUBÍČEK, R. CHMELÍKOVÁ, L. GUMULEC, J. BALVAN, J. PROVAZNÍK, I. KOLÁŘ, R.
Originální název
Self-supervised pretraining for transferable quantitative phase image cell segmentation
Anglický název
Jazyk
en
Originální abstrakt
In this paper, a novel U-Net-based method for robust adherent cell segmentation for quantitative phase microscopy image is designed and optimised. We designed and evaluated four specific post-processing pipelines. To increase the transferability to different cell types, non-deep learning transfer with adjustable parameters is used in the post-processing step. Additionally, we proposed a self-supervised pretraining technique using nonlabelled data, which is trained to reconstruct multiple image distortions and improved the segmentation performance from 0.67 to 0.70 of object-wise intersection over union. Moreover, we publish a new dataset of manually labelled images suitable for this task together with the unlabelled data for self-supervised pretraining.
Anglický abstrakt
Plný text v Digitální knihovně
http://hdl.handle.net/11012/201741
Dokumenty
BibTex
@article{BUT172596, author="Tomáš {Vičar} and Jiří {Chmelík} and Roman {Jakubíček} and Larisa {Chmelíková} and Jaromír {Gumulec} and Jan {Balvan} and Valentine {Provazník} and Radim {Kolář}", title="Self-supervised pretraining for transferable quantitative phase image cell segmentation", annote="In this paper, a novel U-Net-based method for robust adherent cell segmentation for quantitative phase microscopy image is designed and optimised. We designed and evaluated four specific post-processing pipelines. To increase the transferability to different cell types, non-deep learning transfer with adjustable parameters is used in the post-processing step. Additionally, we proposed a self-supervised pretraining technique using nonlabelled data, which is trained to reconstruct multiple image distortions and improved the segmentation performance from 0.67 to 0.70 of object-wise intersection over union. Moreover, we publish a new dataset of manually labelled images suitable for this task together with the unlabelled data for self-supervised pretraining.", address="Optica Publishing Group", chapter="172596", doi="10.1364/BOE.433212", howpublished="online", institution="Optica Publishing Group", number="10", volume="12", year="2021", month="september", pages="6514--6528", publisher="Optica Publishing Group", type="journal article in Web of Science" }