SphereNet: Learning Spherical Representations for Detection and Classification in Omnidirectional Images

Benjamin Coors, Alexandru Paul Condurache, Andreas Geiger

Abstract

Omnidirectional cameras offer great benefits over classical cameras wherever a wide field of view is essential, such as in virtual reality applications or in autonomous robots. Unfortunately, standard convolutional neural networks are not well suited for this scenario as the natural projection surface is a sphere which cannot be unwrapped to a plane without introducing significant distortions, particularly in the polar regions. In this work, we present SphereNet, a novel deep learning framework which encodes invariance against such distortions explicitly into convolutional neural networks. Towards this goal, SphereNet adapts the sampling locations of the convolutional filters, effectively reversing distortions, and wraps the filters around the sphere. By building on regular convolutions, SphereNet enables the transfer of existing perspective convolutional neural network models to the omnidirectional case. We demonstrate the effectiveness of our method on the tasks of image classification and object detection, exploiting two newly created semi-synthetic and real-world omnidirectional datasets.
Original languageEnglish
Title of host publicationComputer Vision -- ECCV 2018
EditorsVittorio Ferrari, Martial Hebert, Cristian Sminchisescu, Yair Weiss
Number of pages17
Volume11213
Place of PublicationCham
PublisherSpringer International Publishing
Publication date05.10.2018
Pages525-541
ISBN (Print)978-3-030-01239-7
ISBN (Electronic)978-3-030-01240-3
DOIs
Publication statusPublished - 05.10.2018
Event15th European Conference on Computer Vision - Munich, Germany
Duration: 08.09.201814.09.2018

Fingerprint

Dive into the research topics of 'SphereNet: Learning Spherical Representations for Detection and Classification in Omnidirectional Images'. Together they form a unique fingerprint.

Cite this