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

Prostate cancer is the most common malignant tumors in men but prostateMagnetic Resonance Imaging (MRI) analysis remains challenging. Besides wholeprostate gland segmentation, the capability to differentiate between the blurryboundary of the Central Gland (CG) and Peripheral Zone (PZ) can lead todifferential diagnosis, since tumor's frequency and severity differ in theseregions. To tackle the prostate zonal segmentation task, we propose a novelConvolutional Neural Network (CNN), called USE-Net, which incorporatesSqueeze-and-Excitation (SE) blocks into U-Net. Especially, the SE blocks areadded after every Encoder (Enc USE-Net) or Encoder-Decoder block (Enc-DecUSE-Net). This study evaluates the generalization ability of CNN-basedarchitectures on three T2-weighted MRI datasets, each one consisting of adifferent number of patients and heterogeneous image characteristics, collectedby different institutions. The following mixed scheme is used fortraining/testing: (i) training on either each individual dataset or multipleprostate MRI datasets and (ii) testing on all three datasets with all possibletraining/testing combinations. USE-Net is compared against threestate-of-the-art CNN-based architectures (i.e., U-Net, pix2pix, and Mixed-ScaleDense Network), along with a semi-automatic continuous max-flow model. Theresults show that training on the union of the datasets generally outperformstraining on each dataset separately, allowing for both intra-/cross-datasetgeneralization. Enc USE-Net shows good overall generalization under anytraining condition, while Enc-Dec USE-Net remarkably outperforms the othermethods when trained on all datasets. These findings reveal that the SE blocks'adaptive feature recalibration provides excellent cross-dataset generalizationwhen testing is performed on samples of the datasets used during training.

USE-Net: incorporating Squeeze-and-Excitation blocks into U-Net for prostate zonal segmentation of multi-institutional MRI datasets