Show simple item record

dc.contributor.author
Gustafsson, Fredrik K.
dc.contributor.author
Danelljan, Martin
dc.contributor.author
Bhat, Goutam
dc.contributor.author
Schön, Thomas B.
dc.contributor.editor
Vedaldi, Andrea
dc.contributor.editor
Bischof, Horst
dc.contributor.editor
Brox, Thomas
dc.contributor.editor
Frahm, Jan-Michael
dc.date.accessioned
2021-01-11T14:34:54Z
dc.date.available
2021-01-10T21:36:35Z
dc.date.available
2021-01-11T14:34:54Z
dc.date.issued
2020
dc.identifier.isbn
978-3-030-58564-8
en_US
dc.identifier.isbn
978-3-030-58565-5
en_US
dc.identifier.issn
0302-9743
dc.identifier.issn
1611-3349
dc.identifier.other
10.1007/978-3-030-58565-5_20
en_US
dc.identifier.uri
http://hdl.handle.net/20.500.11850/460908
dc.description.abstract
While deep learning-based classification is generally tackled using standardized approaches, a wide variety of techniques are employed for regression. In computer vision, one particularly popular such technique is that of confidence-based regression, which entails predicting a confidence value for each input-target pair (x, y). While this approach has demonstrated impressive results, it requires important task-dependent design choices, and the predicted confidences lack a natural probabilistic meaning. We address these issues by proposing a general and conceptually simple regression method with a clear probabilistic interpretation. In our proposed approach, we create an energy-based model of the conditional target density p(y|x), using a deep neural network to predict the un-normalized density from (x, y). This model of p(y|x) is trained by directly minimizing the associated negative log-likelihood, approximated using Monte Carlo sampling. We perform comprehensive experiments on four computer vision regression tasks. Our approach outperforms direct regression, as well as other probabilistic and confidence-based methods. Notably, our model achieves a 2.2% AP improvement over Faster-RCNN for object detection on the COCO dataset, and sets a new state-of-the-art on visual tracking when applied for bounding box estimation. In contrast to confidence-based methods, our approach is also shown to be directly applicable to more general tasks such as age and head-pose estimation. Code is available at https://github.com/fregu856/ebms_regression.
en_US
dc.language.iso
en
en_US
dc.publisher
Springer
en_US
dc.title
Energy-Based Models for Deep Probabilistic Regression
en_US
dc.type
Conference Paper
dc.date.published
2020-11-12
ethz.book.title
Computer Vision – ECCV 2020
en_US
ethz.journal.title
Lecture Notes in Computer Science
ethz.journal.volume
12365
en_US
ethz.journal.abbreviated
LNCS
ethz.pages.start
325
en_US
ethz.pages.end
343
en_US
ethz.event
16th European Conference on Computer Vision (ECCV 2020) (virtual)
en_US
ethz.event.location
Glasgow, United Kingdom
en_US
ethz.event.date
August 23-28, 2020
en_US
ethz.notes
Due to the Coronavirus (COVID-19) the conference was conducted virtually.
en_US
ethz.publication.place
Cham
en_US
ethz.publication.status
published
en_US
ethz.leitzahl
ETH Zürich::00002 - ETH Zürich::00012 - Lehre und Forschung::00007 - Departemente::02140 - Dep. Inf.technologie und Elektrotechnik / Dep. of Inform.Technol. Electrical Eng.::02652 - Institut für Bildverarbeitung / Computer Vision Laboratory::03514 - Van Gool, Luc / Van Gool, Luc
en_US
ethz.leitzahl.certified
ETH Zürich::00002 - ETH Zürich::00012 - Lehre und Forschung::00007 - Departemente::02140 - Dep. Inf.technologie und Elektrotechnik / Dep. of Inform.Technol. Electrical Eng.::02652 - Institut für Bildverarbeitung / Computer Vision Laboratory::03514 - Van Gool, Luc / Van Gool, Luc
en_US
ethz.date.deposited
2021-01-10T21:36:43Z
ethz.source
FORM
ethz.eth
yes
en_US
ethz.availability
Metadata only
en_US
ethz.rosetta.installDate
2021-01-11T14:35:06Z
ethz.rosetta.lastUpdated
2021-02-15T23:07:19Z
ethz.rosetta.versionExported
true
ethz.COinS
ctx_ver=Z39.88-2004&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.atitle=Energy-Based%20Models%20for%20Deep%20Probabilistic%20Regression&rft.jtitle=Lecture%20Notes%20in%20Computer%20Science&rft.date=2020&rft.volume=12365&rft.spage=325&rft.epage=343&rft.issn=0302-9743&1611-3349&rft.au=Gustafsson,%20Fredrik%20K.&Danelljan,%20Martin&Bhat,%20Goutam&Sch%C3%B6n,%20Thomas%20B.&rft.isbn=978-3-030-58564-8&978-3-030-58565-5&rft.genre=proceeding&rft_id=info:doi/10.1007/978-3-030-58565-5_20&rft.btitle=Computer%20Vision%20%E2%80%93%20ECCV%202020
 Search print copy at ETH Library

Files in this item

FilesSizeFormatOpen in viewer

There are no files associated with this item.

Publication type

Show simple item record