EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Robust Automated Tumour Segmentation Network Using 3D Direction-Wise Convolution and Transformer

Ziping Chu, Sonit Singh iD, Arcot Sowmya

Journal of Imaging Informatics in Medicine · 2024

DOI: 10.1007/s10278-024-01131-9 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Imaging Informatics in Medicine, 2024. Verfasst von Ziping Chu und 2 weiteren Autor:innen. Beteiligte Einrichtung: School of Computer Science and Engineering, UNSW Sydney, High St., Kensington, 2052, New South Wales, Australia. Hauptthemen laut MEDLINE: Neural Networks, Computer, Imaging, Three-Dimensional, Brain Neoplasms. Die Arbeit wurde bisher 5-mal zitiert (Crossref) und nennt 48 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: University of New South Wales.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Semantic segmentation of tumours plays a crucial role in fundamental medical image analysis and has a significant impact on cancer diagnosis and treatment planning. UNet and its variants have achieved state-of-the-art results on various 2D and 3D medical image segmentation tasks involving different imaging modalities. Recently, researchers have tried to merge the multi-head self-attention mechanism, as introduced by the Transformer, into U-shaped network structures to enhance the segmentation performance. However, both suffer from limitations that make networks under-perform on voxel-level classification tasks, the Transformer is unable to encode positional information and translation equivariance, while the Convolutional Neural Network lacks global features and dynamic attention. In this work, a new architecture named TCTNet T umour Segmentation with 3D Direction-Wise C onvolution and T ransformer) is introduced, which comprises an encoder utilising a hybrid Transformer-Convolutional Neural Network (CNN) structure and a decoder that incorporates 3D Direction-Wise Convolution. Experimental results show that the proposed hybrid Transformer-CNN network structure obtains better performance than other 3D segmentation networks on the Brain Tumour Segmentation 2021 (BraTS21) dataset. Two more tumour datasets from Medical Segmentation Decathlon are also utilised to test the generalisation ability of the proposed network architecture. In addition, an ablation study was conducted to verify the effectiveness of the designed decoder for the tumour segmentation tasks. The proposed method maintains a competitive segmentation performance while reducing computational effort by 10% in terms of floating-point operations.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Neural Networks, ComputerImaging, Three-DimensionalBrain NeoplasmsHumansAlgorithmsMagnetic Resonance ImagingNeoplasms

Stichwörter der Autor:innen

3D Direction-Wise ConvolutionConvolutional Neural NetworksMedical image segmentationTumour segmentationVision Transformer

Autor:innen und Institutionen

  • School of Computer Science and Engineering, UNSW Sydney, High St., Kensington, 2052, New South Wales, Australia.
  • School of Computer Science and Engineering, UNSW Sydney, High St., Kensington, 2052, New South Wales, Australia. sonit.singh@unsw.edu.au.

ORCID-Profile: Sonit Singh

Förderung

  • University of New South Wales

Zitieren

APA

Chu, Z., Singh, S., & Sowmya, A. (2024). Robust Automated Tumour Segmentation Network Using 3D Direction-Wise Convolution and Transformer. Journal of Imaging Informatics in Medicine. https://doi.org/10.1007/s10278-024-01131-9

Vancouver

Chu Z, Singh S, Sowmya A. Robust Automated Tumour Segmentation Network Using 3D Direction-Wise Convolution and Transformer. Journal of Imaging Informatics in Medicine. 2024. doi:10.1007/s10278-024-01131-9

BibTeX
@article{Chu2024,
  author = {Chu, Ziping and Singh, Sonit and Sowmya, Arcot},
  title = {Robust Automated Tumour Segmentation Network Using 3D Direction-Wise Convolution and Transformer},
  journal = {Journal of Imaging Informatics in Medicine},
  year = {2024},
  doi = {10.1007/s10278-024-01131-9}
}
Kostenlos starten →