Feature Map Augmentation to Improve Rotation Invariance in Convolutional Neural Networks

Research output: A Conference proceeding or a Chapter in BookConference contribution

Abstract

Whilst it is a trivial task for a human vision system to recognize and detect objects with good accuracy, making computer vision algorithms achieve the same feat remains an active area of research. For a human vision system, objects seen once are recognized with high accuracy despite alterations to its appearance by various transformations such as rotations, translations, scale, distortions and occlusion making it a state-of-the-art spatially invariant biological vision system. To make computer algorithms such as Convolutional Neural Networks (CNNs) spatially invariant one popular practice is to introduce variations in the data set through data augmentation. This achieves good results but comes with increased computation cost. In this paper, we address rotation transformation and instead of using data augmentation we propose a novel method that allows CNNs to improve rotation invariance by augmentation of feature maps. This is achieved by creating a rotation transformer layer called Rotation Invariance Transformer (RiT) that can be placed at the output end of a convolution layer. Incoming features are rotated by a given set of rotation parameters which are then passed to the next layer. We test our technique on benchmark CIFAR10 and MNIST datasets in a setting where our RiT layer is placed between the feature extraction and classification layers of the CNN. Our results show promising improvements in the networks ability to be rotation invariant across classes with no increase in model parameters.

Original languageEnglish
Title of host publicationAdvanced Concepts for Intelligent Vision Systems - 20th International Conference, ACIVS 2020, Proceedings
EditorsJacques Blanc-Talon, Patrice Delmas, Wilfried Philips, Dan Popescu, Paul Scheunders
Place of PublicationSwitzerland
PublisherSpringer
Pages348-359
Number of pages12
ISBN (Print)9783030406042
DOIs
Publication statusPublished - 2020
Event20th International Conference on Advanced Concepts for Intelligent Vision Systems - Auckland, New Zealand
Duration: 10 Feb 202014 Feb 2020

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume12002 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference20th International Conference on Advanced Concepts for Intelligent Vision Systems
Abbreviated titleACIVS 2020
CountryNew Zealand
CityAuckland
Period10/02/2014/02/20

    Fingerprint

Cite this

Kumar, D., Sharma, D., & Goecke, R. (2020). Feature Map Augmentation to Improve Rotation Invariance in Convolutional Neural Networks. In J. Blanc-Talon, P. Delmas, W. Philips, D. Popescu, & P. Scheunders (Eds.), Advanced Concepts for Intelligent Vision Systems - 20th International Conference, ACIVS 2020, Proceedings (pp. 348-359). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12002 LNCS). Switzerland: Springer. https://doi.org/10.1007/978-3-030-40605-9_30