by Lebeda, Karel, Hadfield, Simon and Bowden, Richard
Abstract:
In this paper, we address the problem of tracking an unknown object in 3D space. Online 2D tracking often fails for strong out-of-plane rotation which results in considerable changes in appearance beyond those that can be represented by online update strategies. However, by modelling and learning the 3D structure of the object explicitly, such effects are mitigated. To address this, a novel approach is presented, combining techniques from the fields of visual tracking, structure from motion (SfM) and simultaneous localisation and mapping (SLAM). This algorithm is referred to as TMAGIC (Tracking, Modelling And Gaussian-process Inference Combined). At every frame, point and line features are tracked in the image plane and are used, together with their 3D correspondences, to estimate the camera pose. These features are also used to model the 3D shape of the object as a Gaussian process. Tracking determines the trajectories of the object in both the image plane and 3D space, but the approach also provides the 3D object shape. The approach is validated on several video-sequences used in the tracking literature, comparing favourably to state-of-the-art trackers for simple scenes (error reduced by 22 %) with clear advantages in the case of strong out-of-plane rotation, where 2D approaches fail (error reduction of 58 %).
Reference:
2D Or Not 2D: Bridging the Gap Between Tracking and Structure from Motion (Lebeda, Karel, Hadfield, Simon and Bowden, Richard), In Proceedings, Asian Conference on Computer Vision (ACCV) (Cremers, Daniel, Reid, Ian, Saito, Hideo, Yang, Ming-Hsuan, eds.), Springer International Publishing, volume 9006, 2014. (Presentation Slides, Data and Results, Poster)
Bibtex Entry:
@InProceedings{Lebeda14,
Title = {2D Or Not 2D: Bridging the Gap Between Tracking and Structure from Motion},
Author = {Lebeda, Karel and Hadfield, Simon and Bowden, Richard},
Booktitle = {Proceedings, Asian Conference on Computer Vision (ACCV)},
Year = {2014},
Address = {Heidelberg, Germany},
Editor = {Cremers, Daniel and Reid, Ian and Saito, Hideo and Yang, Ming-Hsuan},
Month = {November},
Pages = {642--658},
Publisher = {Springer International Publishing},
Series = {LNCS},
Volume = {9006},
Abstract = {In this paper, we address the problem of tracking an unknown object in 3D space. Online 2D tracking often fails for strong out-of-plane rotation which results in considerable changes in appearance beyond those that can be represented by online update strategies. However, by modelling and learning the 3D structure of the object explicitly, such effects are mitigated. To address this, a novel approach is presented, combining techniques from the fields of visual tracking, structure from motion (SfM) and simultaneous localisation and mapping (SLAM). This algorithm is referred to as TMAGIC (Tracking, Modelling And Gaussian-process Inference Combined). At every frame, point and line features are tracked in the image plane and are used, together with their 3D correspondences, to estimate the camera pose. These features are also used to model the 3D shape of the object as a Gaussian process. Tracking determines the trajectories of the object in both the image plane and 3D space, but the approach also provides the 3D object shape. The approach is validated on several video-sequences used in the tracking literature, comparing favourably to state-of-the-art trackers for simple scenes (error reduced by 22 %) with clear advantages in the case of strong out-of-plane rotation, where 2D approaches fail (error reduction of 58 %).},
Authorship = {34-33-33},
Comment = {<a href="http://cvssp.org/Personal/KarelLebeda/papers/ACCV2014_pres_PRCVC.zip">Presentation Slides</a>, <a href="http://cvssp.org/Personal/KarelLebeda/TMAGIC">Data and Results</a>, <a href="http://cvssp.org/Personal/KarelLebeda/papers/ACCV2014_poster.pdf">Poster</a>},
Day = {1--5},
Doi = {10.1007/978-3-319-16817-3_42},
Gsid = {536887428970724287},
ISBN = {978-3-319-16816-6},
ISSN = {0302-9743},
Keywords = {Visual tracking, Structure from Motion, SLAM, 3D Tracking, Gaussian Process},
Prestige = {international},
Project = {EPSRC EP/I011811/1, BMVA student Travel Grant},
Status = {published},
Url = {http://dx.doi.org/10.1007/978-3-319-16817-3_42},
Venue = {NUS, Singapore},
Year_of_conference = {2014}
}