TY - CHAP
T1 - Bringing 3D models together
T2 - Mining video liaisons in crowdsourced reconstructions
AU - Wang, Ke
AU - Dunn, Enrique
AU - Rodriguez, Mikel
AU - Frahm, Jan Michael
N1 - Publisher Copyright:
© Springer International Publishing AG 2017.
PY - 2017
Y1 - 2017
N2 - The recent advances in large-scale scene modeling have enabled the automatic 3D reconstruction of landmark sites from crowdsourced photo collections. Here, we address the challenge of leveraging crowdsourced video collections to identify connecting visual observations that enable the alignment and subsequent aggregation, of disjoint 3D models. We denote these connecting image sequences as video liaisons and develop a data-driven framework for fully unsupervised extraction and exploitation. Towards this end, we represent video contents in terms of a histogram representation of iconic imagery contained within existing 3D models attained from a photo collection. We then use this representation to efficiently identify and prioritize the analysis of individual videos within a large-scale video collection, in an effort to determine camera motion trajectories connecting different landmarks. Results on crowdsourced data illustrate the efficiency and effectiveness of our proposed approach.
AB - The recent advances in large-scale scene modeling have enabled the automatic 3D reconstruction of landmark sites from crowdsourced photo collections. Here, we address the challenge of leveraging crowdsourced video collections to identify connecting visual observations that enable the alignment and subsequent aggregation, of disjoint 3D models. We denote these connecting image sequences as video liaisons and develop a data-driven framework for fully unsupervised extraction and exploitation. Towards this end, we represent video contents in terms of a histogram representation of iconic imagery contained within existing 3D models attained from a photo collection. We then use this representation to efficiently identify and prioritize the analysis of individual videos within a large-scale video collection, in an effort to determine camera motion trajectories connecting different landmarks. Results on crowdsourced data illustrate the efficiency and effectiveness of our proposed approach.
UR - http://www.scopus.com/inward/record.url?scp=85016059884&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85016059884&partnerID=8YFLogxK
U2 - 10.1007/978-3-319-54190-7_25
DO - 10.1007/978-3-319-54190-7_25
M3 - Chapter
AN - SCOPUS:85016059884
SN - 9783319541891
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 408
EP - 423
BT - Computer Vision - 13th Asian Conference on Computer Vision, ACCV 2016, Revised Selected Papers
A2 - Nishino, Ko
A2 - Lai, Shang-Hong
A2 - Lepetit, Vincent
A2 - Sato, Yoichi
ER -