000268679 001__ 268679
000268679 005__ 20190812204806.0
000268679 037__ $$aCONF
000268679 245__ $$aGeometry Aware Convolutional Filters for Omnidirectional Images Representation
000268679 260__ $$c2019
000268679 269__ $$a2019
000268679 336__ $$aConference Papers
000268679 490__ $$aProceedings of Machine Learning Research$$v97
000268679 520__ $$aDue to their wide field of view, omnidirectional cameras are frequently used by autonomous vehicles, drones and robots for navigation and other computer vision tasks. The images captured by such cameras, are often analyzed and classified with techniques designed for planar images that unfortunately fail to properly handle the native geometry of such images and therefore results in suboptimal performance. In this paper we aim at improving popular deep convolutional neural networks so that they can properly take into account the specific properties of omnidirectional data. In particular we propose an algorithm that adapts convolutional layers, which often serve as a core building block of a CNN, to the properties of omnidirectional images. Thus, our filters have a shape and size that adapt to the location on the omnidirectional image. We show that our method is not limited to spherical surfaces and is able to incorporate the knowledge about any kind of projective geometry inside the deep learning network. As depicted by our experiments, our method outperforms the existing deep neural network techniques for omnidirectional image classification and compression tasks.
000268679 6531_ $$aml-tm
000268679 700__ $$0248106$$aKhasanova, Renata$$g237039
000268679 700__ $$0241061$$aFrossard, Pascal$$g101475
000268679 7112_ $$aICML 2019 - 36th International Conference on Machine Learning$$cLong Beach, CA, USA$$dJan 18-23, 2019
000268679 773__ $$tProceedings of the 36th International Conference on Machine Learning, ICML 2019, 9-15 June 2019, Long Beach, California, USA$$q3351-3359
000268679 8560_ $$falessandra.bianchi@epfl.ch
000268679 85641 $$uhttps://github.com/RenataKh/GAfilters$$yImplementation of the paper
000268679 85641 $$yProceedings$$uhttp://proceedings.mlr.press/v97/
000268679 909C0 $$pLTS4$$mpascal.frossard@epfl.ch$$0252393$$zMarselli, Béatrice$$xU10851
000268679 909CO $$pconf$$pSTI$$ooai:infoscience.epfl.ch:268679
000268679 960__ $$apascal.frossard@epfl.ch
000268679 961__ $$aalessandra.bianchi@epfl.ch
000268679 973__ $$aEPFL$$rREVIEWED
000268679 980__ $$aCONF
000268679 981__ $$aoverwrite