Conference item icon

Conference item

AutoCorrect: Deep inductive alignment of noisy geometric annotations

Abstract:

We propose AutoCorrect, a method to automatically learn object-annotation alignments from a dataset with annotations affected by geometric noise. The method is based on a consistency loss that enables deep neural networks to be trained, given only noisy annotations as input, to correct the annotations. When some noise-free annotations are available, we show that the consistency loss reduces to a stricter self-supervised loss. We also show that the method can implicitly leverage object symmetr...

Expand abstract
Publication status:
Published
Peer review status:
Peer reviewed

Actions


Access Document


Files:
Publication website:
https://bmvc2019.org/programme/detailed-programme/

Authors


More by this author
Institution:
University of Oxford
Department:
Engineering Science
Role:
Author
More by this author
Institution:
University of Oxford
Department:
Engineering Science
Role:
Author
More by this author
Institution:
University of Oxford
Division:
MPLS
Department:
Engineering Science
Oxford college:
New College
Role:
Author
More by this author
Institution:
University of Oxford
Department:
Engineering Science
Role:
Author
Publisher:
British Machine Vision Association Publisher's website
Article number:
152
Publication date:
2020-04-14
Acceptance date:
2019-07-01
Event title:
30th British Machine Vision Conference (BMVC 2019)
Event location:
Cardiff, UK
Event website:
https://bmvc2019.org/
Event start date:
2019-09-09T00:00:00Z
Event end date:
2019-09-12T00:00:00Z
Source identifiers:
1048551
Language:
English
Keywords:
Pubs id:
pubs:1048551
UUID:
uuid:6a76d4b1-acae-4d84-87e5-a4e083f18eed
Local pid:
pubs:1048551
Deposit date:
2019-09-02

Terms of use


Views and Downloads






If you are the owner of this record, you can report an update to it here: Report update to this record

TO TOP