Multi-stage Multi-granularity Focus-Tuned Learning Paradigm for Medical HSI Segmentation

Haichuan Dong; Runjie Zhou; Boxiang Yun; Huihui Zhou; Benyan Zhang; Qingli Li; Yan Wang

doi:10.1007/978-3-031-72111-3_43

Multi-stage Multi-granularity Focus-Tuned Learning Paradigm for Medical HSI Segmentation

Haichuan Dong, Runjie Zhou, Boxiang Yun, Huihui Zhou, Benyan Zhang, Qingli Li, Yan Wang

School of Communication and Electronic Engineering

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

2 Scopus citations

Abstract

Despite significant breakthrough in computational pathology that Medical Hyperspectral Imaging (MHSI) has brought, the asymmetric information in spectral and spatial dimensions pose a primary challenge. In this study, we propose a multi-stage multi-granularity Focus-tuned Learning paradigm for Medical HSI Segmentation. To learn subtle spectral differences while equalizing the spatiospectral feature learning, we design a quadruplet learning pre-training and focus-tuned fine-tuning stages for capturing both disease-level and image-level subtle spectral differences while integrating spatially and spectrally dominant features. We propose an intensifying and weakening strategy throughout all stages. Our method significantly outperforms all competitors in MHSI segmentation, with over 3.5% improvement in DSC. Ablation study further shows our method learns compact spatiospectral features while capturing various levels of spectral differences. Code will be released at https://github.com/DHC233/FL.

Original language	English
Title of host publication	Medical Image Computing and Computer Assisted Intervention – MICCAI 2024 - 27th International Conference, Proceedings
Editors	Marius George Linguraru, Qi Dou, Aasa Feragen, Stamatia Giannarou, Ben Glocker, Karim Lekadir, Julia A. Schnabel
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	456-466
Number of pages	11
ISBN (Print)	9783031721106
DOIs	https://doi.org/10.1007/978-3-031-72111-3_43
State	Published - 2024
Event	27th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2024 - Marrakesh, Morocco Duration: 6 Oct 2024 → 10 Oct 2024

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	15008 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	27th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2024
Country/Territory	Morocco
City	Marrakesh
Period	6/10/24 → 10/10/24

Keywords

Medical hyperspectral images
self-supervised learning

Access to Document

10.1007/978-3-031-72111-3_43

Cite this

Dong, H., Zhou, R., Yun, B., Zhou, H., Zhang, B., Li, Q., & Wang, Y. (2024). Multi-stage Multi-granularity Focus-Tuned Learning Paradigm for Medical HSI Segmentation. In M. G. Linguraru, Q. Dou, A. Feragen, S. Giannarou, B. Glocker, K. Lekadir, & J. A. Schnabel (Eds.), Medical Image Computing and Computer Assisted Intervention – MICCAI 2024 - 27th International Conference, Proceedings (pp. 456-466). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 15008 LNCS). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-031-72111-3_43

Dong, Haichuan ; Zhou, Runjie ; Yun, Boxiang et al. / Multi-stage Multi-granularity Focus-Tuned Learning Paradigm for Medical HSI Segmentation. Medical Image Computing and Computer Assisted Intervention – MICCAI 2024 - 27th International Conference, Proceedings. editor / Marius George Linguraru ; Qi Dou ; Aasa Feragen ; Stamatia Giannarou ; Ben Glocker ; Karim Lekadir ; Julia A. Schnabel. Springer Science and Business Media Deutschland GmbH, 2024. pp. 456-466 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{6be420aad5ee4951b2ee0beb09141618,

title = "Multi-stage Multi-granularity Focus-Tuned Learning Paradigm for Medical HSI Segmentation",

abstract = "Despite significant breakthrough in computational pathology that Medical Hyperspectral Imaging (MHSI) has brought, the asymmetric information in spectral and spatial dimensions pose a primary challenge. In this study, we propose a multi-stage multi-granularity Focus-tuned Learning paradigm for Medical HSI Segmentation. To learn subtle spectral differences while equalizing the spatiospectral feature learning, we design a quadruplet learning pre-training and focus-tuned fine-tuning stages for capturing both disease-level and image-level subtle spectral differences while integrating spatially and spectrally dominant features. We propose an intensifying and weakening strategy throughout all stages. Our method significantly outperforms all competitors in MHSI segmentation, with over 3.5\% improvement in DSC. Ablation study further shows our method learns compact spatiospectral features while capturing various levels of spectral differences. Code will be released at https://github.com/DHC233/FL.",

keywords = "Medical hyperspectral images, self-supervised learning",

author = "Haichuan Dong and Runjie Zhou and Boxiang Yun and Huihui Zhou and Benyan Zhang and Qingli Li and Yan Wang",

note = "Publisher Copyright: {\textcopyright} The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.; 27th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2024 ; Conference date: 06-10-2024 Through 10-10-2024",

year = "2024",

doi = "10.1007/978-3-031-72111-3\_43",

language = "英语",

isbn = "9783031721106",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "456--466",

editor = "Linguraru, \{Marius George\} and Qi Dou and Aasa Feragen and Stamatia Giannarou and Ben Glocker and Karim Lekadir and Schnabel, \{Julia A.\}",

booktitle = "Medical Image Computing and Computer Assisted Intervention – MICCAI 2024 - 27th International Conference, Proceedings",

address = "德国",

}

Dong, H, Zhou, R, Yun, B, Zhou, H, Zhang, B, Li, Q & Wang, Y 2024, Multi-stage Multi-granularity Focus-Tuned Learning Paradigm for Medical HSI Segmentation. in MG Linguraru, Q Dou, A Feragen, S Giannarou, B Glocker, K Lekadir & JA Schnabel (eds), Medical Image Computing and Computer Assisted Intervention – MICCAI 2024 - 27th International Conference, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 15008 LNCS, Springer Science and Business Media Deutschland GmbH, pp. 456-466, 27th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2024, Marrakesh, Morocco, 6/10/24. https://doi.org/10.1007/978-3-031-72111-3_43

Multi-stage Multi-granularity Focus-Tuned Learning Paradigm for Medical HSI Segmentation. / Dong, Haichuan; Zhou, Runjie; Yun, Boxiang et al.
Medical Image Computing and Computer Assisted Intervention – MICCAI 2024 - 27th International Conference, Proceedings. ed. / Marius George Linguraru; Qi Dou; Aasa Feragen; Stamatia Giannarou; Ben Glocker; Karim Lekadir; Julia A. Schnabel. Springer Science and Business Media Deutschland GmbH, 2024. p. 456-466 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 15008 LNCS).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Multi-stage Multi-granularity Focus-Tuned Learning Paradigm for Medical HSI Segmentation

AU - Dong, Haichuan

AU - Zhou, Runjie

AU - Yun, Boxiang

AU - Zhou, Huihui

AU - Zhang, Benyan

AU - Li, Qingli

AU - Wang, Yan

N1 - Publisher Copyright: © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.

PY - 2024

Y1 - 2024

N2 - Despite significant breakthrough in computational pathology that Medical Hyperspectral Imaging (MHSI) has brought, the asymmetric information in spectral and spatial dimensions pose a primary challenge. In this study, we propose a multi-stage multi-granularity Focus-tuned Learning paradigm for Medical HSI Segmentation. To learn subtle spectral differences while equalizing the spatiospectral feature learning, we design a quadruplet learning pre-training and focus-tuned fine-tuning stages for capturing both disease-level and image-level subtle spectral differences while integrating spatially and spectrally dominant features. We propose an intensifying and weakening strategy throughout all stages. Our method significantly outperforms all competitors in MHSI segmentation, with over 3.5% improvement in DSC. Ablation study further shows our method learns compact spatiospectral features while capturing various levels of spectral differences. Code will be released at https://github.com/DHC233/FL.

AB - Despite significant breakthrough in computational pathology that Medical Hyperspectral Imaging (MHSI) has brought, the asymmetric information in spectral and spatial dimensions pose a primary challenge. In this study, we propose a multi-stage multi-granularity Focus-tuned Learning paradigm for Medical HSI Segmentation. To learn subtle spectral differences while equalizing the spatiospectral feature learning, we design a quadruplet learning pre-training and focus-tuned fine-tuning stages for capturing both disease-level and image-level subtle spectral differences while integrating spatially and spectrally dominant features. We propose an intensifying and weakening strategy throughout all stages. Our method significantly outperforms all competitors in MHSI segmentation, with over 3.5% improvement in DSC. Ablation study further shows our method learns compact spatiospectral features while capturing various levels of spectral differences. Code will be released at https://github.com/DHC233/FL.

KW - Medical hyperspectral images

KW - self-supervised learning

UR - https://www.scopus.com/pages/publications/85206883540

U2 - 10.1007/978-3-031-72111-3_43

DO - 10.1007/978-3-031-72111-3_43

M3 - 会议稿件

AN - SCOPUS:85206883540

SN - 9783031721106

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 456

EP - 466

BT - Medical Image Computing and Computer Assisted Intervention – MICCAI 2024 - 27th International Conference, Proceedings

A2 - Linguraru, Marius George

A2 - Dou, Qi

A2 - Feragen, Aasa

A2 - Giannarou, Stamatia

A2 - Glocker, Ben

A2 - Lekadir, Karim

A2 - Schnabel, Julia A.

PB - Springer Science and Business Media Deutschland GmbH

T2 - 27th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2024

Y2 - 6 October 2024 through 10 October 2024

ER -

Dong H, Zhou R, Yun B, Zhou H, Zhang B, Li Q et al. Multi-stage Multi-granularity Focus-Tuned Learning Paradigm for Medical HSI Segmentation. In Linguraru MG, Dou Q, Feragen A, Giannarou S, Glocker B, Lekadir K, Schnabel JA, editors, Medical Image Computing and Computer Assisted Intervention – MICCAI 2024 - 27th International Conference, Proceedings. Springer Science and Business Media Deutschland GmbH. 2024. p. 456-466. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-031-72111-3_43