Please use this identifier to cite or link to this item:
http://hdl.handle.net/1942/42140
Full metadata record
DC Field | Value | Language |
---|---|---|
dc.contributor.author | BECKER, Thijs | - |
dc.contributor.author | ROUSSEAU, Axel-Jan | - |
dc.contributor.author | GEUBBELMANS, Melvin | - |
dc.contributor.author | BURZYKOWSKI, Tomasz | - |
dc.contributor.author | VALKENBORG, Dirk | - |
dc.date.accessioned | 2024-01-16T14:36:01Z | - |
dc.date.available | 2024-01-16T14:36:01Z | - |
dc.date.issued | 2023 | - |
dc.date.submitted | 2024-01-09T19:07:59Z | - |
dc.identifier.citation | Neural Processing Letters, 164 (6) , p. 894 -897 | - |
dc.identifier.uri | http://hdl.handle.net/1942/42140 | - |
dc.description.abstract | Cost volume is widely used to establish correspondences in optical flow estimation. However, when dealing with low-texture and occluded areas, it is difficult to estimate the cost volume correctly. Therefore, we propose a replacement: feature correlation transformer (FCTR), a transformer with self-and cross-attention alternations for obtaining global receptive fields and positional embedding for establishing correspondences. With global context and posi-tional information, FCTR can produce more accurate correspondences for ambiguous areas. Using position-embedded feature allows the removal of the context network; the positional information can be aggregated within ambiguous motion boundaries, and the number of model parameters can be reduced. To speed up network convergence and strengthen robust-ness, we introduce a smooth L1 loss with exponential weights in the pre-training step. At the time of submission, our method achieves competitive performance with all published optical flow methods on both the KITTI-2015 and MPI-Sintel benchmarks. Moreover, it outperforms all optical flow and scene flow methods in KITTI-2015 foreground-region prediction. | - |
dc.language.iso | en | - |
dc.publisher | - | |
dc.subject.other | Humans | - |
dc.subject.other | Decision Trees | - |
dc.subject.other | Random Forest | - |
dc.title | Decision trees and random forests | - |
dc.type | Journal Contribution | - |
dc.identifier.epage | 897 | - |
dc.identifier.issue | 6 | - |
dc.identifier.spage | 894 | - |
dc.identifier.volume | 164 | - |
local.format.pages | 4 | - |
local.bibliographicCitation.jcat | A1 | - |
local.type.refereed | Refereed | - |
local.type.specified | Editorial Material | - |
dc.identifier.doi | 10.1016/j.ajodo.2023.09.011 | - |
dc.identifier.pmid | 38008491 | - |
dc.identifier.isi | WOS:001125458300001 | - |
local.provider.type | PubMed | - |
local.uhasselt.international | yes | - |
item.fullcitation | BECKER, Thijs; ROUSSEAU, Axel-Jan; GEUBBELMANS, Melvin; BURZYKOWSKI, Tomasz & VALKENBORG, Dirk (2023) Decision trees and random forests. In: Neural Processing Letters, 164 (6) , p. 894 -897. | - |
item.contributor | BECKER, Thijs | - |
item.contributor | ROUSSEAU, Axel-Jan | - |
item.contributor | GEUBBELMANS, Melvin | - |
item.contributor | BURZYKOWSKI, Tomasz | - |
item.contributor | VALKENBORG, Dirk | - |
item.fulltext | With Fulltext | - |
item.accessRights | Restricted Access | - |
crisitem.journal.issn | 1370-4621 | - |
crisitem.journal.eissn | 1573-773X | - |
Appears in Collections: | Research publications |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
1-s2.0-S0889540623005188-main.pdf Restricted Access | Published version | 326.22 kB | Adobe PDF | View/Open Request a copy |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.