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 Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

The goal of our study is to segment and quantify brain ventricles in infants with hydrocephalus. The Hydrocephalus is a brain disease in which cerebrospinal fluid accumulates in the ventricles, which expand abnormally. The ventricles then press on other brain tissues, leading to the risk of multiple functional and developmental disorders. Segmenting brain ventricles is necessary for early detection and surgical follow-up. Unfortunately, there are few studies on patients with hydrocephalus and infant ventricles are complex and diverse with limited data. Moreover, using conventional automatic segmentation by atlas and machine learning with handcrafted features is difficult to segment the infant brain ventricles with hydrocephalus because of the above data-specific issues. Here, we propose a deep automatic method based on 2.5D U-Net and transfer learning to segment the infant brain ventricles with hydrocephalus. We apply a network architecture that combines low-level features with high-level features to improve learning efficiency, and to maintain the correlation in the slice direction. The input images of the network are multi-slice images (the target slice image and its neighbor slices). Furthermore, we apply transfer learning using adult datasets to deal with limited data and finetuning in the hydrocephalus infant datasets. In our experiments, our proposed method outperforms conventional methods and improves the DICE from 58% to

Automatic Segmentation of Infant Brain Ventricles with Hydrocephalus in MRI Based on 2.5D U-Net and Transfer Learning