elib
DLR-Header
DLR-Logo -> http://www.dlr.de
DLR Portal Home | Impressum | Datenschutz | Kontakt | English
Schriftgröße: [-] Text [+]

FCDSN-DC: An Accurate and Lightweight Convolutional Neural Network for Stereo Estimation with Depth Completion

Hirner, Dominik und Fraundorfer, Friedrich (2022) FCDSN-DC: An Accurate and Lightweight Convolutional Neural Network for Stereo Estimation with Depth Completion. In: 26th International Conference on Pattern Recognition, ICPR 2022, Seiten 3937-3943. 26th International Conference on Pattern Recognition (ICPR), 2022-08-21 - 2022-08-25, Montreal. doi: 10.1109/ICPR56361.2022.9956175. ISBN 978-1-66549-062-7. ISSN 1051-4651.

[img] PDF
333kB

Offizielle URL: https://www.computer.org/csdl/proceedings-article/icpr/2022/09956175/1IHq3gwrlEk

Kurzfassung

We propose an accurate and lightweight convolutional neural network for stereo estimation with depth completion. We name this method fully-convolutional deformable similarity network with depth completion (FCDSN-DC). This method extends FC-DCNN by improving the feature extractor, adding a network structure for training highly accurate similarity functions and a network structure for filling inconsistent disparity estimates. The whole method consists of three parts. The first part consists of fully-convolutional densely connected layers that computes expressive features of rectified image pairs. The second part of our network learns highly accurate similarity functions between this learned features. It consists of densely-connected convolution layers with a deformable convolution block at the end to further improve the accuracy of the results. After this step an initial disparity map is created and the left-right consistency check is performed in order to remove inconsistent points. The last part of the network then uses this input together with the corresponding left RGB image in order to train a network that fills in the missing measurements. Consistent depth estimations are gathered around invalid points and are parsed together with the RGB points into a shallow CNN network structure in order to recover the missing values. We evaluate our method on challenging real world indoor and outdoor scenes, in particular Middlebury, KITTI and ETH3D were it produces competitive results. We furthermore show that this method generalizes well and is well suited for many applications without the need of further training. The code of our full framework is available at: https://github.com/thedodo/FCDSN-DC

elib-URL des Eintrags:https://elib.dlr.de/190991/
Dokumentart:Konferenzbeitrag (Vortrag)
Titel:FCDSN-DC: An Accurate and Lightweight Convolutional Neural Network for Stereo Estimation with Depth Completion
Autoren:
AutorenInstitution oder E-Mail-AdresseAutoren-ORCID-iDORCID Put Code
Hirner, DominikInstitute for Computer Graphics and VisionNICHT SPEZIFIZIERTNICHT SPEZIFIZIERT
Fraundorfer, Friedrichfriedrich.fraundorfer (at) dlr.deNICHT SPEZIFIZIERTNICHT SPEZIFIZIERT
Datum:2022
Erschienen in:26th International Conference on Pattern Recognition, ICPR 2022
Referierte Publikation:Ja
Open Access:Ja
Gold Open Access:Nein
In SCOPUS:Ja
In ISI Web of Science:Nein
DOI:10.1109/ICPR56361.2022.9956175
Seitenbereich:Seiten 3937-3943
ISSN:1051-4651
ISBN:978-1-66549-062-7
Status:veröffentlicht
Stichwörter:lightweight convolutional neural network, depth completion, fully-convolutional deformable similarity network
Veranstaltungstitel:26th International Conference on Pattern Recognition (ICPR)
Veranstaltungsort:Montreal
Veranstaltungsart:internationale Konferenz
Veranstaltungsbeginn:21 August 2022
Veranstaltungsende:25 August 2022
HGF - Forschungsbereich:Luftfahrt, Raumfahrt und Verkehr
HGF - Programm:Raumfahrt
HGF - Programmthema:Erdbeobachtung
DLR - Schwerpunkt:Raumfahrt
DLR - Forschungsgebiet:R EO - Erdbeobachtung
DLR - Teilgebiet (Projekt, Vorhaben):R - Optische Fernerkundung, V - Digitaler Atlas 2.0
Standort: Oberpfaffenhofen
Institute & Einrichtungen:Institut für Methodik der Fernerkundung > Photogrammetrie und Bildanalyse
Hinterlegt von: Knickl, Sabine
Hinterlegt am:06 Dez 2022 17:44
Letzte Änderung:24 Apr 2024 20:52

Nur für Mitarbeiter des Archivs: Kontrollseite des Eintrags

Blättern
Suchen
Hilfe & Kontakt
Informationen
electronic library verwendet EPrints 3.3.12
Gestaltung Webseite und Datenbank: Copyright © Deutsches Zentrum für Luft- und Raumfahrt (DLR). Alle Rechte vorbehalten.