Detail publikace

3D Dense-U-Net for MRI brain tissue segmentation

Originální název

3D Dense-U-Net for MRI brain tissue segmentation

Anglický název

3D Dense-U-Net for MRI brain tissue segmentation

Jazyk

en

Originální abstrakt

This paper presents a fully automatic method for 3D segmentation of brain tissue on MRI scans using modern deep learning approach and proposes 3D Dense-U-Net neural network architecture using densely connected layers. In contrast with many previous methods, our approach is capable of precise segmentation without any preprocessing of the input image and achieved accuracy 99.70 percent on testing data which outperformed human expert results. The architecture proposed in this paper can also be easily applied to any project already using U-net network as a segmentation algorithm to enhance its results. Implementation was done in Keras on Tensorflow backend and complete source-code was released online.

Anglický abstrakt

This paper presents a fully automatic method for 3D segmentation of brain tissue on MRI scans using modern deep learning approach and proposes 3D Dense-U-Net neural network architecture using densely connected layers. In contrast with many previous methods, our approach is capable of precise segmentation without any preprocessing of the input image and achieved accuracy 99.70 percent on testing data which outperformed human expert results. The architecture proposed in this paper can also be easily applied to any project already using U-net network as a segmentation algorithm to enhance its results. Implementation was done in Keras on Tensorflow backend and complete source-code was released online.

BibTex


@inproceedings{BUT148982,
  author="Martin {Kolařík} and Radim {Burget} and Václav {Uher} and Malay Kishore {Dutta}",
  title="3D Dense-U-Net for MRI brain tissue segmentation",
  annote="This paper presents a fully automatic method for 3D segmentation of brain tissue on MRI scans using modern deep learning approach and proposes 3D Dense-U-Net neural network architecture using densely connected layers. In contrast with many previous methods, our approach is capable of precise segmentation without any preprocessing of the input image and achieved accuracy 99.70 percent on testing data which outperformed human expert results. The architecture proposed in this paper can also be easily applied to any project already using U-net network as a segmentation algorithm to enhance its results. Implementation was done in Keras on Tensorflow backend and complete source-code was released online.",
  address="IEEE",
  booktitle="Proceedings of the 2018 41st International Conference on Telecommunications and Signal Processing (TSP)",
  chapter="148982",
  doi="10.1109/TSP.2018.8441508",
  howpublished="online",
  institution="IEEE",
  year="2018",
  month="july",
  pages="237--240",
  publisher="IEEE",
  type="conference paper"
}