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

Magnetic resonance imaging (MRI) data is heterogeneous due to differences indevice manufacturers, scanning protocols, and inter-subject variability. Aconventional way to mitigate MR image heterogeneity is to apply preprocessingtransformations such as anatomy alignment, voxel resampling, signal intensityequalization, image denoising, and localization of regions of interest.Although a preprocessing pipeline standardizes image appearance, its influenceon the quality of image segmentation and on other downstream tasks in deepneural networks has never been rigorously studied.  We conduct experiments on three publicly available datasets and evaluate theeffect of different preprocessing steps in intra- and inter-dataset trainingscenarios. Our results demonstrate that most popular standardization steps addno value to the network performance; moreover, preprocessing can hamper modelperformance. We suggest that image intensity normalization approaches do notcontribute to model accuracy because of the reduction of signal variance withimage standardization. Finally, we show that the contribution ofskull-stripping in data preprocessing is almost negligible if measured in termsof estimated tumor volume.  We show that the only essential transformation for accurate deep learninganalysis is the unification of voxel spacing across the dataset. In contrast,inter-subjects anatomy alignment in the form of non-rigid atlas registration isnot necessary and intensity equalization steps (denoising, bias-fieldcorrection and histogram matching) do not improve models' performance. Thestudy code is accessible onlinehttps://github.com/MedImAIR/brain-mri-processing-pipeline

Negligible effect of brain MRI data preprocessing for tumor segmentation