elib
DLR-Header
DLR-Logo -> http://www.dlr.de
DLR Portal Home | Imprint | Privacy Policy | Contact | Deutsch
Fontsize: [-] Text [+]

FC-DCNN: A densely connected neural network for stereo estimation

Hirner, Dominik and Fraundorfer, Friedrich (2021) FC-DCNN: A densely connected neural network for stereo estimation. ICPR 2020, 10.-15. Jan. 2021, Milan, Italy.

[img] PDF
9MB

Abstract

We propose a novel lightweight network for stereo estimation. Our network consists of a fully-convolutional densely connected neural network (FC-DCNN) that computes matching costs between rectified image pairs. Our FC-DCNN method learns expressive features and performs some simple but effective postprocessing steps. The densely connected layer structure connects the output of each layer to the input of each subsequent layer. This network structure and the fact that we do not use any fullyconnected layers or 3D convolutions leads to a very lightweight network. The output of this network is used in order to calculate matching costs and create a cost-volume. Instead of using time and memory-inefficient cost-aggregation methods such as semiglobal matching or conditional random fields in order to improve the result, we rely on filtering techniques, namely median filter and guided filter. By computing a left-right consistency check we get rid of inconsistent values. Afterwards we use a watershed foreground-background segmentation on the disparity image with removed inconsistencies. This mask is then used to refine the final prediction. We show that our method works well for both challenging indoor and outdoor scenes by evaluating it on the Middlebury, KITTI and ETH3D benchmarks respectively. Our full framework is available at https://github.com/thedodo/FCDCNN

Item URL in elib:https://elib.dlr.de/138907/
Document Type:Conference or Workshop Item (Poster)
Title:FC-DCNN: A densely connected neural network for stereo estimation
Authors:
AuthorsInstitution or Email of AuthorsAuthor's ORCID iD
Hirner, DominikInstitute for Computer Graphics and VisionUNSPECIFIED
Fraundorfer, Friedrichfriedrich.fraundorfer (at) dlr.deUNSPECIFIED
Date:2021
Refereed publication:Yes
Open Access:Yes
Gold Open Access:No
In SCOPUS:No
In ISI Web of Science:No
Status:Accepted
Keywords:3D vision; Deep learning; Neural networks
Event Title:ICPR 2020
Event Location:Milan, Italy
Event Type:international Conference
Event Dates:10.-15. Jan. 2021
HGF - Research field:Aeronautics, Space and Transport
HGF - Program:Space
HGF - Program Themes:Earth Observation
DLR - Research area:Raumfahrt
DLR - Program:R EO - Earth Observation
DLR - Research theme (Project):R - Vorhaben hochauflösende Fernerkundungsverfahren (old)
Location: Oberpfaffenhofen
Institutes and Institutions:Remote Sensing Technology Institute > Photogrammetry and Image Analysis
Deposited By: Knickl, Sabine
Deposited On:01 Dec 2020 15:56
Last Modified:01 Mar 2021 03:00

Repository Staff Only: item control page

Browse
Search
Help & Contact
Information
electronic library is running on EPrints 3.3.12
Copyright © 2008-2017 German Aerospace Center (DLR). All rights reserved.