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

The extension of convolutional neural networks (CNNs) to non-Euclideangeometries has led to multiple frameworks for studying manifolds. Many of thosemethods have shown design limitations resulting in poor modelling of long-rangeassociations, as the generalisation of convolutions to irregular surfaces isnon-trivial. Motivated by the success of attention-modelling in computervision, we translate convolution-free vision transformer approaches to surfacedata, to introduce a domain-agnostic architecture to study any surface dataprojected onto a spherical manifold. Here, surface patching is achieved byrepresenting spherical data as a sequence of triangular patches, extracted froma subdivided icosphere. A transformer model encodes the sequence of patches viasuccessive multi-head self-attention layers while preserving the sequenceresolution. We validate the performance of the proposed Surface VisionTransformer (SiT) on the task of phenotype regression from cortical surfacemetrics derived from the Developing Human Connectome Project (dHCP).Experiments show that the SiT generally outperforms surface CNNs, whileperforming comparably on registered and unregistered data. Analysis oftransformer attention maps offers strong potential to characterise subtlecognitive developmental patterns.

Surface Vision Transformers: Attention-Based Modelling applied to Cortical Analysis