Efficient Bayesian deep inversion

Higham, C. F. , Johnson, S. , Radwell, N., Padgett, M. J. and Murray-Smith, R. (2023) Efficient Bayesian deep inversion. Journal of Computational Dynamics, (doi: 10.3934/jcd.2023014) (Early Online Publication)

[img] Text
309898.pdf - Accepted Version
Available under License Creative Commons Attribution.



We develop a deep learning method to enhance sensor detection for depth prediction. Our novel system combines sensor hardware and Bayesian inference to solve the underlying inverse problem, recovering depth from measurements. The hardware comprises single sensor non-scanning time-of-flight laser detection with synchronised video to produce a 3D depth map. The Bayesian framework provides depth prediction with uncertainty quantification. A conditional generator-discriminator adversarial network is adapted to learn a compact representation of the scene that recovers 3D depth at 30 Hz using a large training set. We transfer the network to a real hardware system and compare with ground truth depth information. Our novel synthesis of hardware and machine learning technologies addresses the important challenge of providing accurate absolute depth prediction at video rate with efficient and cost-effective non-scanning laser detection technology. This flexible and compact system has many exciting applications for autonomous vehicles, drones and wearable technology.

Item Type:Articles
Additional Information:The authors would like to thank and acknowledge the financial support from the 39 Engineering and Physical Sciences Research Council (EPSRC) (QuantIC Nos. 40 EP/M01326X/1, EP/T00097X/1 and Looking and Listening in Complex Media No. 41 EP/S026444/1); and also the H2020 European Research Council (ERC) (TWISTS, 42 No. 340507).
Status:Early Online Publication
Glasgow Author(s) Enlighten ID:Murray-Smith, Professor Roderick and Higham, Dr Catherine and Radwell, Dr Neal and Johnson, Dr Steven and Padgett, Professor Miles
Authors: Higham, C. F., Johnson, S., Radwell, N., Padgett, M. J., and Murray-Smith, R.
College/School:College of Science and Engineering > School of Computing Science
College of Science and Engineering > School of Physics and Astronomy
Journal Name:Journal of Computational Dynamics
Publisher:American Institute of Mathematical Sciences
ISSN (Online):2158-2505
Published Online:23 December 2023
Copyright Holders:Copyright © 2023 American Institute of Mathematical Sciences
First Published:First published in Journal of Computational Dynamics 2023
Publisher Policy:Reproduced in accordance with the publisher copyright policy

University Staff: Request a correction | Enlighten Editors: Update this record

Project CodeAward NoProject NamePrincipal InvestigatorFunder's NameFunder RefLead Dept
190841UK Quantum Technology Hub in Enhanced Quantum ImagingMiles PadgettEngineering and Physical Sciences Research Council (EPSRC)EP/M01326X/1P&S - Physics & Astronomy
305029Looking and Listening in Complex Media - UDRADaniele FaccioEngineering and Physical Sciences Research Council (EPSRC)EP/S026444/1P&S - Physics & Astronomy
305567QuantIC - The UK Quantum Technoogy Hub in Quantum Enhanced ImagingMiles PadgettEngineering and Physical Sciences Research Council (EPSRC)EP/T00097X/1P&S - Physics & Astronomy
169636Twists and more: the complex shape of lightMiles PadgettEuropean Research Council (ERC)340507P&S - Physics & Astronomy