Repository logo
 

Probabilistic Future Prediction for Video Scene Understanding

Simple item page
cam.orpheus.counter29*
dc.contributor.authorHu, Anthony
dc.contributor.authorCotter, Fergal
dc.contributor.authorMohan, Nikhil
dc.contributor.authorGurau, Corina
dc.contributor.authorKendall, Alex
dc.date.accessioned2021-10-20T23:30:09Z
dc.date.available2021-10-20T23:30:09Z
dc.description.abstractWe present a novel deep learning architecture for probabilistic future prediction from video. We predict the future semantics, geometry and motion of complex real-world urban scenes and use this representation to control an autonomous vehicle. This work is the first to jointly predict ego-motion, static scene, and the motion of dynamic agents in a probabilistic manner, which allows sampling consistent, highly probable futures from a compact latent space. Our model learns a representation from RGB video with a spatio-temporal convolutional module. The learned representation can be explicitly decoded to future semantic segmentation, depth, and optical flow, in addition to being an input to a learnt driving policy. To model the stochasticity of the future, we introduce a conditional variational approach which minimises the divergence between the present distribution (what could happen given what we have seen) and the future distribution (what we observe actually happens). During inference, diverse futures are generated by sampling from the present distribution.
dc.description.sponsorshipToshiba Europe, grant G100453
dc.identifier.doi10.17863/CAM.77127
dc.identifier.urihttps://www.repository.cam.ac.uk/handle/1810/329681
dc.language.isoeng
dc.rightsAll rights reserved
dc.rights.urihttp://www.rioxx.net/licenses/all-rights-reserved
dc.titleProbabilistic Future Prediction for Video Scene Understanding
dc.typeConference Object
dcterms.dateAccepted2020-07-03
pubs.conference-finish-date2020-08-28
pubs.conference-nameEuropean Conference on Computer Vision (ECCV)
pubs.conference-start-date2020-08-23
rioxxterms.licenseref.startdate2020-07-03
rioxxterms.licenseref.urihttp://www.rioxx.net/licenses/all-rights-reserved
rioxxterms.typeConference Paper/Proceeding/Abstract
rioxxterms.versionAM
rioxxterms.versionofrecord10.17863/CAM.77127

Files

Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
ECCV2020___Probabilistic_Future_Prediction_for_Video_Scene_Understanding.pdf
Size:
4.94 MB
Format:
Adobe Portable Document Format
Description:
Accepted version
Licence
http://www.rioxx.net/licenses/all-rights-reserved
Download
License bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
DepositLicenceAgreementv2.1.pdf
Size:
150.9 KB
Format:
Adobe Portable Document Format
Download

Collections

University of Cambridge Research Outputs (Articles and Conferences)