Advanced search
1 file | 4.84 MB

Self-learning voxel-based multi-camera occlusion maps for 3D reconstruction

Author
Organization
Abstract
The quality of a shape-from-silhouettes 3D reconstruction technique strongly depends on the completeness of the silhouettes from each of the cameras. Static occlusion, due to e.g. furniture, makes reconstruction difficult, as we assume no prior knowledge concerning shape and size of occluding objects in the scene. In this paper we present a self-learning algorithm that is able to build an occlusion map for each camera from a voxel perspective. This information is then used to determine which cameras need to be evaluated when reconstructing the 3D model at every voxel in the scene. We show promising results in a multi-camera setup with seven cameras where the object is significantly better reconstructed compared to the state of the art methods, despite the occluding object in the center of the room.
Keywords
Occlusion Detection, Visual hull, Multi-camera, Self-learning

Downloads

  • visapp.pdf
    • full text
    • |
    • open access
    • |
    • PDF
    • |
    • 4.84 MB

Citation

Please use this url to cite or link to this publication:

Chicago
Slembrouck, Maarten, Dimitri Van Cauwelaert, David Van Hamme, Dirk Van Haerenborgh, Peter Van Hese, Peter Veelaert, and Wilfried Philips. 2014. “Self-learning Voxel-based Multi-camera Occlusion Maps for 3D Reconstruction.” In PROCEEDINGS OF THE 2014 9TH INTERNATIONAL CONFERENCE ON COMPUTER VISION, THEORY AND APPLICATIONS (VISAPP 2014), VOL 2 , ed. Sebastiano Battiato and José Braz, 502–509. SCITEPRESS.
APA
Slembrouck, M., Van Cauwelaert, D., Van Hamme, D., Van Haerenborgh, D., Van Hese, P., Veelaert, P., & Philips, W. (2014). Self-learning voxel-based multi-camera occlusion maps for 3D reconstruction. In S. Battiato & J. Braz (Eds.), PROCEEDINGS OF THE 2014 9TH INTERNATIONAL CONFERENCE ON COMPUTER VISION, THEORY AND APPLICATIONS (VISAPP 2014), VOL 2 (pp. 502–509). Presented at the 9th International Conference on Computer Vision Theory and Applications (VISAPP) , SCITEPRESS.
Vancouver
1.
Slembrouck M, Van Cauwelaert D, Van Hamme D, Van Haerenborgh D, Van Hese P, Veelaert P, et al. Self-learning voxel-based multi-camera occlusion maps for 3D reconstruction. In: Battiato S, Braz J, editors. PROCEEDINGS OF THE 2014 9TH INTERNATIONAL CONFERENCE ON COMPUTER VISION, THEORY AND APPLICATIONS (VISAPP 2014), VOL 2 . SCITEPRESS; 2014. p. 502–9.
MLA
Slembrouck, Maarten, Dimitri Van Cauwelaert, David Van Hamme, et al. “Self-learning Voxel-based Multi-camera Occlusion Maps for 3D Reconstruction.” PROCEEDINGS OF THE 2014 9TH INTERNATIONAL CONFERENCE ON COMPUTER VISION, THEORY AND APPLICATIONS (VISAPP 2014), VOL 2 . Ed. Sebastiano Battiato & José Braz. SCITEPRESS, 2014. 502–509. Print.
@inproceedings{4229490,
  abstract     = {The quality of a shape-from-silhouettes 3D reconstruction technique strongly depends on the completeness of the silhouettes from each of the cameras. Static occlusion, due to e.g. furniture, makes reconstruction difficult, as we assume no prior knowledge concerning shape and size of occluding objects in the scene. In this paper we present a self-learning algorithm that is able to build an occlusion map for each camera from a voxel perspective. This information is then used to determine which cameras need to be evaluated when reconstructing the 3D model at every voxel in the scene. We show promising results in a multi-camera setup with seven cameras where the object is significantly better reconstructed compared to the state of the art methods, despite the occluding object in the center of the room.},
  author       = {Slembrouck, Maarten and Van Cauwelaert, Dimitri and Van Hamme, David and Van Haerenborgh, Dirk and Van Hese, Peter and Veelaert, Peter and Philips, Wilfried},
  booktitle    = {PROCEEDINGS OF THE 2014 9TH INTERNATIONAL CONFERENCE ON COMPUTER VISION, THEORY AND APPLICATIONS (VISAPP 2014), VOL 2 },
  editor       = {Battiato, Sebastiano and Braz, Jos{\'e}},
  isbn         = {978-9-8975-8133-5},
  keyword      = {Occlusion Detection,Visual hull,Multi-camera,Self-learning},
  language     = {eng},
  location     = {Lisbon, Portugal},
  pages        = {502--509},
  publisher    = {SCITEPRESS},
  title        = {Self-learning voxel-based multi-camera occlusion maps for 3D reconstruction},
  year         = {2014},
}

Web of Science
Times cited: