Self-supervised pretraining for transferable quantitative phase
image cell segmentation

J 2021

Self-supervised pretraining for transferable quantitative phase image cell segmentation

VIČAR, Tomáš; Jiri CHEMELIK; Roman JAKUBICEK; Larisa CHMELIKOVA; Jaromír GUMULEC et. al.

Basic information

Original name

Self-supervised pretraining for transferable quantitative phase image cell segmentation

Authors

VIČAR, Tomáš (203 Czech Republic, belonging to the institution); Jiri CHEMELIK (203 Czech Republic); Roman JAKUBICEK (203 Czech Republic); Larisa CHMELIKOVA (203 Czech Republic); Jaromír GUMULEC (203 Czech Republic, belonging to the institution); Jan BALVAN (203 Czech Republic, belonging to the institution); Ivo PROVAZNÍK (203 Czech Republic) and Radim KOLAR (203 Czech Republic, guarantor)

Edition

BIOMEDICAL OPTICS EXPRESS, WASHINGTON, OPTICAL SOC AMER, 2021, 2156-7085

Other information

Language

English

Type of outcome

Article in a journal

Field of Study

30224 Radiology, nuclear medicine and medical imaging

Country of publisher

United States of America

Confidentiality degree

is not subject to a state or trade secret

References:

URL

Impact factor

Impact factor: 3.562

RIV identification code

RIV/00216224:14110/21:00123919

Organization unit

Faculty of Medicine

DOI

https://doi.org/10.1364/BOE.433212

UT WoS

000703871700005

EID Scopus

2-s2.0-85115888948

Keywords in English

Self-supervised pretraining; transferable quantitative phase image cell segmentation

Abstract

In the original language

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. (c) 2021 Optical Society of America under the terms of the OSA Open Access Publishing Agreement

Links

LM2018140, research and development project

Name: e-Infrastruktura CZ (Acronym: e-INFRA CZ)

Investor: Ministry of Education, Youth and Sports of the CR

Cite

VIČAR, Tomáš; Jiri CHEMELIK; Roman JAKUBICEK; Larisa CHMELIKOVA; Jaromír GUMULEC; Jan BALVAN; Ivo PROVAZNÍK and Radim KOLAR. Self-supervised pretraining for transferable quantitative phase image cell segmentation. BIOMEDICAL OPTICS EXPRESS. WASHINGTON: OPTICAL SOC AMER, 2021, vol. 12, No 10, p. 6514-6528. ISSN 2156-7085. Available from: https://doi.org/10.1364/BOE.433212.

@article{1826219,
   author = {Vičar, Tomáš and Chemelik, Jiri and Jakubicek, Roman and Chmelikova, Larisa and Gumulec, Jaromír and Balvan, Jan and Provazník, Ivo and Kolar, Radim},
   article_location = {WASHINGTON},
   article_number = {10},
   doi = {https://doi.org/10.1364/BOE.433212},
   keywords = {Self-supervised pretraining; transferable quantitative phase image cell segmentation},
   language = {eng},
   issn = {2156-7085},
   journal = {BIOMEDICAL OPTICS EXPRESS},
   title = {Self-supervised pretraining for transferable quantitative phase image cell segmentation},
   url = {https://www.osapublishing.org/boe/fulltext.cfm?uri=boe-12-10-6514&id=459853},
   volume = {12},
   year = {2021}
}

TY  - JOUR
ID  - 1826219
AU  - Vičar, Tomáš - Chemelik, Jiri - Jakubicek, Roman - Chmelikova, Larisa - Gumulec, Jaromír - Balvan, Jan - Provazník, Ivo - Kolar, Radim
PY  - 2021
TI  - Self-supervised pretraining for transferable quantitative phase image cell segmentation
JF  - BIOMEDICAL OPTICS EXPRESS
VL  - 12
IS  - 10
SP  - 6514-6528
EP  - 6514-6528
PB  - OPTICAL SOC AMER
SN  - 21567085
KW  - Self-supervised pretraining
KW  - transferable quantitative phase image cell segmentation
UR  - https://www.osapublishing.org/boe/fulltext.cfm?uri=boe-12-10-6514&id=459853
N2  - 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. (c) 2021 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
ER  -

VIČAR, Tomáš; Jiri CHEMELIK; Roman JAKUBICEK; Larisa CHMELIKOVA; Jaromír GUMULEC; Jan BALVAN; Ivo PROVAZNÍK and Radim KOLAR. Self-supervised pretraining for transferable quantitative phase image cell segmentation. \textit{BIOMEDICAL OPTICS EXPRESS}. WASHINGTON: OPTICAL SOC AMER, 2021, vol.~12, No~10, p.~6514-6528. ISSN~2156-7085. Available from: https://doi.org/10.1364/BOE.433212.

Přehled o publikaci