AI Paper Atlas

Research map / Architectures and efficiency / CNN design

Achieving Rotational Invariance with Bessel-Convolutional Neural Networks

Valentin Delchevalerie, Adrien Bibal, Benoit Frenay, Alexandre Mayer

NeurIPS 2021 · Architectures and efficiency › CNN design · approach: Compact CNN architectures · 7 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use Bessel convolutions instead of group convolutionsB-CNN · Outex feature map equivariancefeature map difference (%): 10.64% -> 1.77%Improved
use Bessel convolutions instead of standard convolutionsB-CNN · Outex feature map equivariancefeature map difference (%): 5.73% -> 1.77%Improved
add random rotation data augmentation to training (ablation)all models (B-CNN, G-CNN, CNN) · MNIST, Outex, brain MRItest accuracyImproved
B-CNN vs G-CNN without data augmentationB-CNN · MNISTtest accuracyImproved
B-CNN vs G-CNN without data augmentationB-CNN · Outextest accuracyNo clear effect
B-CNN vs G-CNN with data augmentationB-CNN · Outextest accuracy: less than 5% accuracyImproved
low-data setting, 75% of training set removedB-CNN vs G-CNN · Outex low datatest accuracy: a little bit less than 15%Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy