Publication detail

Denoise pre-training for segmentation neural networks

KOLAŘÍK, M.

Original Title

Denoise pre-training for segmentation neural networks

Type

conference paper

Language

English

Original Abstract

This paper proposes a method for pre-training segmentation neural networks on small datasets using unlabelled training data with added noise. The pre-training process helps the network with initial better weights settings for the training itself and also augments the training dataset when dealing with small labelled datasets especially in medical imaging. The experiment comparing results of pre-trained and not pre-trained networks on MRI brain segmentation task has shown that the denoise pre-training helps the network with faster training convergence without overfitting and achieving better results in all compared metrics even on very small datasets.

Keywords

deep learning; denoising; neural network; pre-training; segmentation

Authors

KOLAŘÍK, M.

Released

25. 4. 2019

Publisher

Vysoké učení technické v Brně, Fakulta elektrotechniky a komunikačních technologií

Location

Brno

ISBN

978-80-214-5735-5

Book

Proceedings of the 25th Conference STUDENT EEICT 2019

Pages from

739

Pages to

744

Pages count

5

BibTex

@inproceedings{BUT157996,
  author="Martin {Kolařík}",
  title="Denoise pre-training for segmentation neural networks",
  booktitle="Proceedings of the 25th Conference STUDENT EEICT 2019",
  year="2019",
  pages="739--744",
  publisher="Vysoké učení technické v Brně, Fakulta elektrotechniky a komunikačních technologií",
  address="Brno",
  isbn="978-80-214-5735-5"
}