Denoise pre-training for segmentation neural networks

Denoise pre-training for segmentation neural networks

Paper in proceedings (conference paper)

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.

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.

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

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

KOLAŘÍK, M.

2020

25.04.2019

Vysoké učení technické v Brně, Fakulta elektrotechniky a komunikačních technologií

Brno

978-80-214-5735-5

Proceedings of the 25th Conference STUDENT EEICT 2019

739

744

5

EEICT_Denoise_pretraining