Reference : A case study on the impact of masking moving objects on the camera pose regression wi...
Scientific congresses, symposiums and conference proceedings : Paper published in a book
Engineering, computing & technology : Computer science
Computational Sciences
http://hdl.handle.net/10993/41613
A case study on the impact of masking moving objects on the camera pose regression with CNNs
English
Cimarelli, Claudio mailto [University of Luxembourg > Interdisciplinary Centre for Security, Reliability and Trust (SNT) >]
Cazzato, Dario mailto [University of Luxembourg > Interdisciplinary Centre for Security, Reliability and Trust (SNT) >]
Olivares Mendez, Miguel Angel mailto [University of Luxembourg > Interdisciplinary Centre for Security, Reliability and Trust (SNT) > >]
Voos, Holger mailto [University of Luxembourg > Interdisciplinary Centre for Security, Reliability and Trust (SNT) > Engineering Research Unit]
25-Nov-2019
2019 16th IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS)
1--8
Yes
No
International
2019 16th IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS)
from 18-09-2019 to 21-09-2019
IEEE
Taipei
Taiwan
[en] Cameras ; Neural Networks ; Feature Extraction ; Training ; Visualization ; Image Segmentation ; Simultaneous localization and mapping
[en] Robot self-localization is essential for operating autonomously in open environments. When cameras are the main source of information for retrieving the pose, numerous challenges are posed by the presence of dynamic objects, due to occlusion and continuous changes in the appearance. Recent research on global localization methods focused on using a single (or multiple) Convolutional Neural Network (CNN) to estimate the 6 Degrees of Freedom (6-DoF) pose directly from a monocular camera image. In contrast with the classical approaches using engineered feature detector, CNNs are usually more robust to environmental changes in light and to occlusions in outdoor scenarios. This paper contains an attempt to empirically demonstrate the ability of CNNs to ignore dynamic elements, such as pedestrians or cars, through learning. For this purpose, we pre-process a dataset for pose localization with an object segmentation network, masking potentially moving objects. Hence, we compare the pose regression CNN trained and/or tested on the set of masked images and the original one. Experimental results show that the performances of the two training approaches are similar, with a slight reduction of the error when hiding occluding objects from the views.
http://hdl.handle.net/10993/41613
10.1109/AVSS.2019.8909904
https://ieeexplore.ieee.org/abstract/document/8909904

File(s) associated to this reference

Fulltext file(s):

FileCommentaryVersionSizeAccess
Open access
DLAM_maskposenet.pdfAuthor preprint13.6 MBView/Open

Bookmark and Share SFX Query

All documents in ORBilu are protected by a user license.