Enhancing MRI Brain Tumor Segmentation with an Additional Classification Network

TL;DR

This paper introduces a novel training approach for MRI brain tumor segmentation by adding a classification branch to a deep neural network, significantly improving segmentation accuracy on the BraTS 2020 dataset.

Contribution

The paper presents a new training method that incorporates an additional classification network branch to enhance segmentation performance in brain tumor MRI analysis.

Findings

Achieved Dice scores of 78.43%, 89.99%, and 84.22% on BraTS validation set.

Improved segmentation accuracy over baseline models.

Demonstrated effectiveness of the classification branch in end-to-end training.

Abstract

Brain tumor segmentation plays an essential role in medical image analysis. In recent studies, deep convolution neural networks (DCNNs) are extremely powerful to tackle tumor segmentation tasks. We propose in this paper a novel training method that enhances the segmentation results by adding an additional classification branch to the network. The whole network was trained end-to-end on the Multimodal Brain Tumor Segmentation Challenge (BraTS) 2020 training dataset. On the BraTS's validation set, it achieved an average Dice score of 78.43%, 89.99%, and 84.22% respectively for the enhancing tumor, the whole tumor, and the tumor core.