HNF-Netv2 for Brain Tumor Segmentation using multi-modal MR Imaging

TL;DR

HNF-Netv2 enhances brain tumor segmentation by incorporating semantic discrimination blocks, achieving high accuracy on BraTS 2021 and winning the RSNA 2021 challenge.

Contribution

This work introduces HNF-Netv2 with semantic discrimination blocks, improving upon prior HNF-Net for multi-modal MR brain tumor segmentation.

Findings

Achieved Dice scores of approximately 0.88, 0.87, and 0.92 for different tumor regions.

Attained Hausdorff distances of 8.92, 16.25, and 4.49 for the tumor regions.

Secured 8th place in the RSNA 2021 Brain Tumor AI Challenge.

Abstract

In our previous work, $i.e.$, HNF-Net, high-resolution feature representation and light-weight non-local self-attention mechanism are exploited for brain tumor segmentation using multi-modal MR imaging. In this paper, we extend our HNF-Net to HNF-Netv2 by adding inter-scale and intra-scale semantic discrimination enhancing blocks to further exploit global semantic discrimination for the obtained high-resolution features. We trained and evaluated our HNF-Netv2 on the multi-modal Brain Tumor Segmentation Challenge (BraTS) 2021 dataset. The result on the test set shows that our HNF-Netv2 achieved the average Dice scores of 0.878514, 0.872985, and 0.924919, as well as the Hausdorff distances ($95\%$) of 8.9184, 16.2530, and 4.4895 for the enhancing tumor, tumor core, and whole tumor, respectively. Our method won the RSNA 2021 Brain Tumor AI Challenge Prize (Segmentation Task), which ranks 8th out of all 1250 submitted results.