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

Humans and animals recognize objects irrespective of the beholder's point ofview, which may drastically change their appearances. Artificial patternrecognizers also strive to achieve this, e.g., through translational invariancein convolutional neural networks (CNNs). However, both CNNs and visiontransformers (ViTs) perform very poorly on rotated inputs. Here we presentartificial mental rotation (AMR), a novel deep learning paradigm for dealingwith in-plane rotations inspired by the neuro-psychological concept of mentalrotation. Our simple AMR implementation works with all common CNN and ViTarchitectures. We test it on ImageNet, Stanford Cars, and Oxford Pet. With atop-1 error (averaged across datasets and architectures) of $0.743$, AMRoutperforms the current state of the art (rotational data augmentation, averagetop-1 error of $0.626$) by $19\%$. We also easily transfer a trained AMR moduleto a downstream task to improve the performance of a pre-trained semanticsegmentation model on rotated CoCo from $32.7$ to $55.2$ IoU.

Efficient Rotation Invariance in Deep Neural Networks through Artificial Mental Rotation