Multimodal 3D Brain Tumor Segmentation with Adversarial Training and Conditional Random Field

Lan  Jiang; Yuchao Zheng; Miao Yu; Haiqing Zhang; Fatemah Aladwani; Alessandro Perelli

doi:10.48550/arXiv.2411.14418

Multimodal 3D Brain Tumor Segmentation with Adversarial Training and Conditional Random Field

Lan Jiang, Yuchao Zheng, Miao Yu, Haiqing Zhang, Fatemah Aladwani, Alessandro Perelli (Lead / Corresponding author)

Biomedical Engineering

Research output: Working paper/Preprint › Preprint

15 Downloads (Pure)

Abstract

Accurate brain tumor segmentation remains a challenging task due to structural complexity and great individual differences of gliomas. Leveraging the pre-eminent detail resilience of CRF and spatial feature extraction capacity of V-net, we propose a multimodal 3D Volume Generative Adversarial Network (3D-vGAN) for precise segmentation. The model utilizes Pseudo-3D for V-net improvement, adds conditional random field after generator and use original image as supplemental guidance. Results, using the BraTS-2018 [12] dataset, show that 3D-vGAN outperforms classical segmentation models, including U-net, Gan, FCN and 3D V-net, reaching specificity over 99.8%.

Original language	English
Publisher	arXiv
Number of pages	13
DOIs	https://doi.org/10.48550/arXiv.2411.14418
Publication status	Published - 21 Nov 2024

Keywords

Multimodal Segmentation
Generative Adversarial Network
Brain tumor

Access to Document

10.48550/arXiv.2411.14418Licence: CC BY

Pre-PrintSubmitted manuscript, 2.59 MBLicence: CC BY

1 Conference contribution

Multimodal 3D Brain Tumor Segmentation with Adversarial Training and Conditional Random Field
Jiang, L., Zheng, Y., Yu, M., Zhang, H., Aladwani, F. & Perelli, A. (Lead / Corresponding author), 24 Jul 2024, Medical Image Understanding and Analysis: 28th Annual Conference, MIUA 2024, Proceedings. Yap, M. H., Kendrick, C., Behera, A., Cootes, T. & Zwiggelaar, R. (eds.). Springer , Vol. 14859. p. 68-80 13 p. (Lecture Notes in Computer Science; vol. 14859 LNCS).
Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Cite this

@techreport{87d7ce342962493682e9941f5041bf50,

title = "Multimodal 3D Brain Tumor Segmentation with Adversarial Training and Conditional Random Field",

abstract = "Accurate brain tumor segmentation remains a challenging task due to structural complexity and great individual differences of gliomas. Leveraging the pre-eminent detail resilience of CRF and spatial feature extraction capacity of V-net, we propose a multimodal 3D Volume Generative Adversarial Network (3D-vGAN) for precise segmentation. The model utilizes Pseudo-3D for V-net improvement, adds conditional random field after generator and use original image as supplemental guidance. Results, using the BraTS-2018 [12] dataset, show that 3D-vGAN outperforms classical segmentation models, including U-net, Gan, FCN and 3D V-net, reaching specificity over 99.8%.",

keywords = "Multimodal Segmentation, Generative Adversarial Network, Brain tumor",

author = "Lan Jiang and Yuchao Zheng and Miao Yu and Haiqing Zhang and Fatemah Aladwani and Alessandro Perelli",

year = "2024",

month = nov,

day = "21",

doi = "10.48550/arXiv.2411.14418",

language = "English",

publisher = "arXiv",

type = "WorkingPaper",

institution = "arXiv",

}

TY - UNPB

T1 - Multimodal 3D Brain Tumor Segmentation with Adversarial Training and Conditional Random Field

AU - Jiang, Lan

AU - Zheng, Yuchao

AU - Yu, Miao

AU - Zhang, Haiqing

AU - Aladwani, Fatemah

AU - Perelli, Alessandro

PY - 2024/11/21

Y1 - 2024/11/21

N2 - Accurate brain tumor segmentation remains a challenging task due to structural complexity and great individual differences of gliomas. Leveraging the pre-eminent detail resilience of CRF and spatial feature extraction capacity of V-net, we propose a multimodal 3D Volume Generative Adversarial Network (3D-vGAN) for precise segmentation. The model utilizes Pseudo-3D for V-net improvement, adds conditional random field after generator and use original image as supplemental guidance. Results, using the BraTS-2018 [12] dataset, show that 3D-vGAN outperforms classical segmentation models, including U-net, Gan, FCN and 3D V-net, reaching specificity over 99.8%.

AB - Accurate brain tumor segmentation remains a challenging task due to structural complexity and great individual differences of gliomas. Leveraging the pre-eminent detail resilience of CRF and spatial feature extraction capacity of V-net, we propose a multimodal 3D Volume Generative Adversarial Network (3D-vGAN) for precise segmentation. The model utilizes Pseudo-3D for V-net improvement, adds conditional random field after generator and use original image as supplemental guidance. Results, using the BraTS-2018 [12] dataset, show that 3D-vGAN outperforms classical segmentation models, including U-net, Gan, FCN and 3D V-net, reaching specificity over 99.8%.

KW - Multimodal Segmentation

KW - Generative Adversarial Network

KW - Brain tumor

U2 - 10.48550/arXiv.2411.14418

DO - 10.48550/arXiv.2411.14418

M3 - Preprint

BT - Multimodal 3D Brain Tumor Segmentation with Adversarial Training and Conditional Random Field

PB - arXiv

ER -

Multimodal 3D Brain Tumor Segmentation with Adversarial Training and Conditional Random Field

Abstract

Keywords

Access to Document

Fingerprint

Research output

Multimodal 3D Brain Tumor Segmentation with Adversarial Training and Conditional Random Field

Cite this