Robustcaps: a transformation-robust capsule network for image classification

TL;DR

RobustCaps is a transformation-robust capsule network that employs group-equivariant convolutions and a context-normalized routing algorithm to improve image classification under geometric transformations, outperforming existing models on standard datasets.

Contribution

Introduces RobustCaps, a novel capsule network with group-equivariant convolutions and a context-normalized routing algorithm for transformation-robust image classification.

Findings

Achieves state-of-the-art accuracy on CIFAR-10, FashionMNIST, and CIFAR-100 under rotations and translations.

Outperforms baseline capsule and convolutional neural networks on transformation-robust tasks.

Demonstrates the effectiveness of group-equivariant convolutions and context normalization in capsule networks.

Abstract

Geometric transformations of the training data as well as the test data present challenges to the use of deep neural networks to vision-based learning tasks. In order to address this issue, we present a deep neural network model that exhibits the desirable property of transformation-robustness. Our model, termed RobustCaps, uses group-equivariant convolutions in an improved capsule network model. RobustCaps uses a global context-normalised procedure in its routing algorithm to learn transformation-invariant part-whole relationships within image data. This learning of such relationships allows our model to outperform both capsule and convolutional neural network baselines on transformation-robust classification tasks. Specifically, RobustCaps achieves state-of-the-art accuracies on CIFAR-10, FashionMNIST, and CIFAR-100 when the images in these datasets are subjected to train and test-time rotations and translations.