Ensemble of Multi-sized FCNs to Improve White Matter Lesion Segmentation

TL;DR

This paper introduces a two-stage multi-sized FCN ensemble approach with a novel activation function to improve white-matter lesion segmentation accuracy in brain MRI scans, effectively handling lesion size variability.

Contribution

The study proposes a novel multi-sized FCN ensemble framework with a new activation function, enhancing segmentation performance for lesions of varying sizes.

Findings

Improved Dice Similarity Coefficient on WMH data

Effective handling of lesion size variability

Superior performance over existing methods

Abstract

In this paper, we develop a two-stage neural network solution for the challenging task of white-matter lesion segmentation. To cope with the vast vari- ability in lesion sizes, we sample brain MR scans with patches at three differ- ent dimensions and feed them into separate fully convolutional neural networks (FCNs). In the second stage, we process large and small lesion separately, and use ensemble-nets to combine the segmentation results generated from the FCNs. A novel activation function is adopted in the ensemble-nets to improve the segmen- tation accuracy measured by Dice Similarity Coefficient. Experiments on MICCAI 2017 White Matter Hyperintensities (WMH) Segmentation Challenge data demonstrate that our two-stage-multi-sized FCN approach, as well as the new activation function, are effective in capturing white-matter lesions in MR images.