TY - GEN
T1 - A Light-Weight Interpretable Model for Nuclei Detection and Weakly-Supervised Segmentation
AU - Zhang, Yixiao
AU - Kortylewski, Adam
AU - Liu, Qing
AU - Park, Seyoun
AU - Green, Benjamin
AU - Engle, Elizabeth
AU - Almodovar, Guillermo
AU - Walk, Ryan
AU - Soto-Diaz, Sigfredo
AU - Taube, Janis
AU - Szalay, Alex
AU - Yuille, Alan
N1 - Publisher Copyright:
© 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.
PY - 2022
Y1 - 2022
N2 - The field of computational pathology has witnessed great advancements since deep neural networks have been widely applied. These networks usually require large numbers of annotated data to train vast parameters. However, it takes significant effort to annotate a large histo-pathology dataset. We introduce a light-weight and interpretable model for nuclei detection and weakly-supervised segmentation. It only requires annotations on isolated nucleus, rather than on all nuclei in the dataset. Besides, it is a generative compositional model that first locates parts of nucleus, then learns the spatial correlation of the parts to further locate the nucleus. This process brings interpretability in its prediction. Empirical results on an in-house dataset show that in detection, the proposed method achieved comparable or better performance than its deep network counterparts, especially when the annotated data is limited. It also outperforms popular weakly-supervised segmentation methods. The proposed method could be an alternative solution for the data-hungry problem of deep learning methods.
AB - The field of computational pathology has witnessed great advancements since deep neural networks have been widely applied. These networks usually require large numbers of annotated data to train vast parameters. However, it takes significant effort to annotate a large histo-pathology dataset. We introduce a light-weight and interpretable model for nuclei detection and weakly-supervised segmentation. It only requires annotations on isolated nucleus, rather than on all nuclei in the dataset. Besides, it is a generative compositional model that first locates parts of nucleus, then learns the spatial correlation of the parts to further locate the nucleus. This process brings interpretability in its prediction. Empirical results on an in-house dataset show that in detection, the proposed method achieved comparable or better performance than its deep network counterparts, especially when the annotated data is limited. It also outperforms popular weakly-supervised segmentation methods. The proposed method could be an alternative solution for the data-hungry problem of deep learning methods.
KW - Nuclei detection and segmentation
KW - Weakly-supervised
UR - http://www.scopus.com/inward/record.url?scp=85138829187&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85138829187&partnerID=8YFLogxK
U2 - 10.1007/978-3-031-16961-8_15
DO - 10.1007/978-3-031-16961-8_15
M3 - Conference contribution
AN - SCOPUS:85138829187
SN - 9783031169601
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 145
EP - 155
BT - Medical Optical Imaging and Virtual Microscopy Image Analysis - 1st International Workshop, MOVI 2022, Held in Conjunction with MICCAI 2022, Proceedings
A2 - Huo, Yuankai
A2 - Millis, Bryan A.
A2 - Zhou, Yuyin
A2 - Wang, Xiangxue
A2 - Harrison, Adam P.
A2 - Xu, Ziyue
PB - Springer Science and Business Media Deutschland GmbH
T2 - 1st International Workshop on Medical Optical Imaging and Virtual Microscopy Image Analysis, MOVI 2022, held in conjunction with the 25th International Conference on Medical Imaging and Computer Assisted Intervention, MICCAI 2022
Y2 - 18 September 2022 through 18 September 2022
ER -