TY - GEN
T1 - SCOPE
T2 - 5th Workshop on GRaphs in biomedicAl Image anaLysis Satellite event at MICCAI, GRAIL 2023 and 1st Cell Detection from Cell-Tissue Interaction challenge in MICCAI, OCELOT 2023 Held in Conjunction with International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2023
AU - Yeganeh, Yousef
AU - Güvercin, Göktuğ
AU - Xiao, Rui
AU - Abuzer, Amr
AU - Adeli, Ehsan
AU - Farshad, Azade
AU - Navab, Nassir
N1 - Publisher Copyright:
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.
PY - 2024
Y1 - 2024
N2 - Although the preservation of shape continuity and physiological anatomy is a natural assumption in the segmentation of medical images, it is often neglected by deep learning methods that mostly aim for the statistical modeling of input data as pixels rather than interconnected structures. In biological structures, however, organs are not separate entities; for example, in reality, a severed vessel is an indication of an underlying problem, but traditional segmentation models are not designed to strictly enforce the continuity of anatomy, potentially leading to inaccurate medical diagnoses. To address this issue, we propose a graph-based approach that enforces the continuity and connectivity of anatomical topology in medical images. Our method encodes the continuity of shapes as a graph constraint, ensuring that the network’s predictions maintain this continuity. We evaluate our method on three public benchmarks of retinal vessel segmentation and one neuronal structure segmentation benchmark, showing significant improvements in connectivity metrics compared to previous works while getting better or on-par performance on segmentation metrics.
AB - Although the preservation of shape continuity and physiological anatomy is a natural assumption in the segmentation of medical images, it is often neglected by deep learning methods that mostly aim for the statistical modeling of input data as pixels rather than interconnected structures. In biological structures, however, organs are not separate entities; for example, in reality, a severed vessel is an indication of an underlying problem, but traditional segmentation models are not designed to strictly enforce the continuity of anatomy, potentially leading to inaccurate medical diagnoses. To address this issue, we propose a graph-based approach that enforces the continuity and connectivity of anatomical topology in medical images. Our method encodes the continuity of shapes as a graph constraint, ensuring that the network’s predictions maintain this continuity. We evaluate our method on three public benchmarks of retinal vessel segmentation and one neuronal structure segmentation benchmark, showing significant improvements in connectivity metrics compared to previous works while getting better or on-par performance on segmentation metrics.
UR - http://www.scopus.com/inward/record.url?scp=85188739768&partnerID=8YFLogxK
U2 - 10.1007/978-3-031-55088-1_1
DO - 10.1007/978-3-031-55088-1_1
M3 - Conference contribution
AN - SCOPUS:85188739768
SN - 9783031550874
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 3
EP - 13
BT - Graphs in Biomedical Image Analysis, and Overlapped Cell on Tissue Dataset for Histopathology - 5th MICCAI Workshop, GRAIL 2023 and 1st MICCAI Challenge, OCELOT 2023, Held in Conjunction with MICCAI 2023, Proceedings
A2 - Ahmadi, Seyed-Ahmad
A2 - Pereira, Sérgio
PB - Springer Science and Business Media Deutschland GmbH
Y2 - 8 October 2023 through 12 October 2023
ER -