AI2Seg: A Method and Tool for AI-based Annotation Inspection of Biomedical Instance Segmentation Datasets

Marcel P. Schilling; Lukas Klinger; Ulrike Schumacher; Svenja Schmelzer; Miguel Bordallo López; Britta Nestler; Markus Reischl

doi:10.1109/EMBC40787.2023.10341074

AI²Seg: A Method and Tool for AI-based Annotation Inspection of Biomedical Instance Segmentation Datasets

Marcel P. Schilling^*, Lukas Klinger, Ulrike Schumacher, Svenja Schmelzer, Miguel Bordallo López, Britta Nestler, Markus Reischl

^*Corresponding author for this work

Not published at VTT

Research output: Chapter in Book/Report/Conference proceeding › Conference article in proceedings › Scientific › peer-review

1 Citation (Scopus)

Abstract

In biomedical engineering, deep neural networks are commonly used for the diagnosis and assessment of diseases through the interpretation of medical images. The effectiveness of these networks relies heavily on the availability of annotated datasets for training. However, obtaining noise-free and consistent annotations from experts, such as pathologists, radiologists, and biologists, remains a significant challenge. One common task in clinical practice and biological imaging applications is instance segmentation. Though, there is currently a lack of methods and open-source tools for the automated inspection of biomedical instance segmentation datasets concerning noisy annotations. To address this issue, we propose a novel deep learning-based approach for inspecting noisy annotations and provide an accompanying software implementation, AI2Seg, to facilitate its use by domain experts. The performance of the proposed algorithm is demonstrated on the medical MoNuSeg dataset and the biological LIVECell dataset.

Original language	English
Title of host publication	2023 45th Annual International Conference of the IEEE Engineering in Medicine and Biology Conference, EMBC 2023 - Proceedings
Publisher	IEEE Institute of Electrical and Electronic Engineers
Number of pages	6
ISBN (Electronic)	9798350324471
DOIs	https://doi.org/10.1109/EMBC40787.2023.10341074
Publication status	Published - 2023
MoE publication type	A4 Article in a conference publication
Event	45th Annual International Conference of the IEEE Engineering in Medicine and Biology Conference, EMBC 2023 - Sydney, Australia Duration: 24 Jul 2023 → 27 Jul 2023

Conference

Conference	45th Annual International Conference of the IEEE Engineering in Medicine and Biology Conference, EMBC 2023
Country/Territory	Australia
City	Sydney
Period	24/07/23 → 27/07/23

Funding

*This work was funded by the Future Fields initative ”Screening Platform for Personalized Oncology” of the KIT strategy of excellence. The developed methods are also provided within the KNMFi program (no. 43.31.01) of the Helmholtz association. We also thank the Karlsruhe House of Young Scientists for funding the scientific exchange in terms of the “Networking Grant” that was useful in obtaining the results. This work was supported by the HoreKa Supercomputer through the Ministry of Science, Research, and the Arts Baden-Württemberg.

Keywords

Humans
Algorithms
Bioengineering
Biomedical Engineering
Health Personnel
Neural Networks, Computer

Access to Document

10.1109/EMBC40787.2023.10341074

Cite this

Schilling, M. P., Klinger, L., Schumacher, U., Schmelzer, S., López, M. B., Nestler, B., & Reischl, M. (2023). AI²Seg: A Method and Tool for AI-based Annotation Inspection of Biomedical Instance Segmentation Datasets. In 2023 45th Annual International Conference of the IEEE Engineering in Medicine and Biology Conference, EMBC 2023 - Proceedings IEEE Institute of Electrical and Electronic Engineers. https://doi.org/10.1109/EMBC40787.2023.10341074

@inproceedings{3517dad37e5b42ecb1df5e9bc2bdf8dc,

title = "AI2Seg: A Method and Tool for AI-based Annotation Inspection of Biomedical Instance Segmentation Datasets",

abstract = "In biomedical engineering, deep neural networks are commonly used for the diagnosis and assessment of diseases through the interpretation of medical images. The effectiveness of these networks relies heavily on the availability of annotated datasets for training. However, obtaining noise-free and consistent annotations from experts, such as pathologists, radiologists, and biologists, remains a significant challenge. One common task in clinical practice and biological imaging applications is instance segmentation. Though, there is currently a lack of methods and open-source tools for the automated inspection of biomedical instance segmentation datasets concerning noisy annotations. To address this issue, we propose a novel deep learning-based approach for inspecting noisy annotations and provide an accompanying software implementation, AI2Seg, to facilitate its use by domain experts. The performance of the proposed algorithm is demonstrated on the medical MoNuSeg dataset and the biological LIVECell dataset.",

keywords = "Humans, Algorithms, Bioengineering, Biomedical Engineering, Health Personnel, Neural Networks, Computer",

author = "Schilling, {Marcel P.} and Lukas Klinger and Ulrike Schumacher and Svenja Schmelzer and L{\'o}pez, {Miguel Bordallo} and Britta Nestler and Markus Reischl",

note = "Publisher Copyright: {\textcopyright} 2023 IEEE.; 45th Annual International Conference of the IEEE Engineering in Medicine and Biology Conference, EMBC 2023 ; Conference date: 24-07-2023 Through 27-07-2023",

year = "2023",

doi = "10.1109/EMBC40787.2023.10341074",

language = "English",

booktitle = "2023 45th Annual International Conference of the IEEE Engineering in Medicine and Biology Conference, EMBC 2023 - Proceedings",

publisher = "IEEE Institute of Electrical and Electronic Engineers",

address = "United States",

}

Schilling, MP, Klinger, L, Schumacher, U, Schmelzer, S, López, MB, Nestler, B & Reischl, M 2023, AI²Seg: A Method and Tool for AI-based Annotation Inspection of Biomedical Instance Segmentation Datasets. in 2023 45th Annual International Conference of the IEEE Engineering in Medicine and Biology Conference, EMBC 2023 - Proceedings. IEEE Institute of Electrical and Electronic Engineers, 45th Annual International Conference of the IEEE Engineering in Medicine and Biology Conference, EMBC 2023, Sydney, Australia, 24/07/23. https://doi.org/10.1109/EMBC40787.2023.10341074

AI²Seg: A Method and Tool for AI-based Annotation Inspection of Biomedical Instance Segmentation Datasets. / Schilling, Marcel P.; Klinger, Lukas; Schumacher, Ulrike et al.
2023 45th Annual International Conference of the IEEE Engineering in Medicine and Biology Conference, EMBC 2023 - Proceedings. IEEE Institute of Electrical and Electronic Engineers, 2023.

Research output: Chapter in Book/Report/Conference proceeding › Conference article in proceedings › Scientific › peer-review

TY - GEN

T1 - AI2Seg

T2 - 45th Annual International Conference of the IEEE Engineering in Medicine and Biology Conference, EMBC 2023

AU - Schilling, Marcel P.

AU - Klinger, Lukas

AU - Schumacher, Ulrike

AU - Schmelzer, Svenja

AU - López, Miguel Bordallo

AU - Nestler, Britta

AU - Reischl, Markus

PY - 2023

Y1 - 2023

N2 - In biomedical engineering, deep neural networks are commonly used for the diagnosis and assessment of diseases through the interpretation of medical images. The effectiveness of these networks relies heavily on the availability of annotated datasets for training. However, obtaining noise-free and consistent annotations from experts, such as pathologists, radiologists, and biologists, remains a significant challenge. One common task in clinical practice and biological imaging applications is instance segmentation. Though, there is currently a lack of methods and open-source tools for the automated inspection of biomedical instance segmentation datasets concerning noisy annotations. To address this issue, we propose a novel deep learning-based approach for inspecting noisy annotations and provide an accompanying software implementation, AI2Seg, to facilitate its use by domain experts. The performance of the proposed algorithm is demonstrated on the medical MoNuSeg dataset and the biological LIVECell dataset.

AB - In biomedical engineering, deep neural networks are commonly used for the diagnosis and assessment of diseases through the interpretation of medical images. The effectiveness of these networks relies heavily on the availability of annotated datasets for training. However, obtaining noise-free and consistent annotations from experts, such as pathologists, radiologists, and biologists, remains a significant challenge. One common task in clinical practice and biological imaging applications is instance segmentation. Though, there is currently a lack of methods and open-source tools for the automated inspection of biomedical instance segmentation datasets concerning noisy annotations. To address this issue, we propose a novel deep learning-based approach for inspecting noisy annotations and provide an accompanying software implementation, AI2Seg, to facilitate its use by domain experts. The performance of the proposed algorithm is demonstrated on the medical MoNuSeg dataset and the biological LIVECell dataset.

KW - Humans

KW - Algorithms

KW - Bioengineering

KW - Biomedical Engineering

KW - Health Personnel

KW - Neural Networks, Computer

UR - http://www.scopus.com/inward/record.url?scp=85179642678&partnerID=8YFLogxK

U2 - 10.1109/EMBC40787.2023.10341074

DO - 10.1109/EMBC40787.2023.10341074

M3 - Conference article in proceedings

C2 - 38083322

AN - SCOPUS:85179642678

BT - 2023 45th Annual International Conference of the IEEE Engineering in Medicine and Biology Conference, EMBC 2023 - Proceedings

PB - IEEE Institute of Electrical and Electronic Engineers

Y2 - 24 July 2023 through 27 July 2023

ER -

Schilling MP, Klinger L, Schumacher U, Schmelzer S, López MB, Nestler B et al. AI²Seg: A Method and Tool for AI-based Annotation Inspection of Biomedical Instance Segmentation Datasets. In 2023 45th Annual International Conference of the IEEE Engineering in Medicine and Biology Conference, EMBC 2023 - Proceedings. IEEE Institute of Electrical and Electronic Engineers. 2023 doi: 10.1109/EMBC40787.2023.10341074