English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Manual delineation of tumor regions from magnetic resonance (MR) images istime-consuming, requires an expert, and is prone to human error. In recentyears, deep learning models have been the go-to approach for the segmentationof brain tumors. U-Net and its' variants for semantic segmentation of medicalimages have achieved good results in the literature. However, U-Net and its'variants tend to over-segment tumor regions and may not accurately segment thetumor edges. The edges of the tumor are as important as the tumor regions foraccurate diagnosis, surgical precision, and treatment planning. In the proposedwork, the authors aim to extract edges from the ground truth using aderivative-like filter followed by edge reconstruction to obtain an edge groundtruth in addition to the brain tumor ground truth. Utilizing both groundtruths, the author studies several U-Net and its' variant architectures withand without tumor edges ground truth as a target along with the tumor groundtruth for brain tumor segmentation. The author used the BraTS2020 benchmarkdataset to perform the study and the results are tabulated for the dice andHausdorff95 metrics. The mean and median metrics are calculated for the wholetumor (WT), tumor core (TC), and enhancing tumor (ET) regions. Compared to thebaseline U-Net and its variants, the models that learned edges along with thetumor regions performed well in core tumor regions in both training andvalidation datasets. The improved performance of edge-trained models trained onbaseline models like U-Net and V-Net achieved performance similar to baselinestate-of-the-art models like Swin U-Net and hybrid MR-U-Net. The edge-targettrained models are capable of generating edge maps that can be useful fortreatment planning. Additionally, for further explainability of the results,the activation map generated by the hybrid MR-U-Net has been studied.

Integrating Edges into U-Net Models with Explainable Activation Maps for Brain Tumor Segmentation using MR Images