UCL Discovery
UCL home » Library Services » Electronic resources » UCL Discovery

A New Rotation-Invariant Deep Network for 3D Object Recognition

Zhang, Y; Lu, Z; Xue, J-H; Liao, Q; (2019) A New Rotation-Invariant Deep Network for 3D Object Recognition. In: Karam, Lina J and Mei, Tao and Wu, Feng, (eds.) Proceedings of the 2019 IEEE International Conference on Multimedia and Expo (ICME). IEEE Green open access

[img]
Preview
Text
YachiZhang-ICME2019.pdf - Accepted version

Download (616kB) | Preview

Abstract

When inputs are rotated, most 3D convolutional neural networks (CNNs) will have their performance much dropped, especially for those models with voxelized input of 3D objects. The newly proposed Spherical CNNS, with the concept of the rotation-equivariant spherical correlation, aims to achieve rotation invariance. Inspired by this, we propose a new rotation-invariant deep network to recognize rotated 3D objects. Specifically, we adopt the spherical representation and the spherical correlation S^2 layer of Spherical CNNs, for their capacity of representing 3D objects and rotation equivariance. In the meantime, we improve the computational efficiency and expressiveness of Spherical CNNs, by replacing its time-consuming and depth-limited SO(3) layer with a PointNet-style network architecture. Hence our proposed network can maintain the equivariance as the network grows deeper while substantially reducing its runtime, leading to a much better efficiency and expressiveness of rotation-invariant representation. Experimental results show that our network performs better than or comparable to the state-of-the-art methods in the ModelNet40 classification challenge.

Type: Proceedings paper
Title: A New Rotation-Invariant Deep Network for 3D Object Recognition
ISBN: 978-1-5386-9552-4
Open access status: An open access version is available from UCL Discovery
DOI: 10.1109/icme.2019.00277
Publisher version: https://doi.org/10.1109/ICME.2019.00277
Language: English
Additional information: This version is the author accepted manuscript. For information on re-use, please refer to the publisher’s terms and conditions.
Keywords: Three-dimensional displays, Correlation, Two-dimensional displays, Feature extraction, Solid modeling, Convolution, Training, Deep learning, Object recognition, Rotation invariant, 3D representation
UCL classification: UCL
UCL > Provost and Vice Provost Offices
UCL > Provost and Vice Provost Offices > UCL BEAMS
UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of Maths and Physical Sciences
UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of Maths and Physical Sciences > Dept of Statistical Science
URI: https://discovery.ucl.ac.uk/id/eprint/10079666
Downloads since deposit
75Downloads
Download activity - last month
Download activity - last 12 months
Downloads by country - last 12 months

Archive Staff Only

View Item View Item