A 3-year funded PhD position for an
excellent student is available in the field of Computer Graphics and
Geometric Modeling at the University of Strasbourg (France), ICube
Lab (http://icube.unistra.fr).
Title: Reconstruction and visualization of digitized 3D models from scans and photographs.
Host research team:
Computer Graphics and Geometry group (http://icube-igg.unistra.fr/en)
Computer Graphics and Geometry group (http://icube-igg.unistra.fr/en)
Advisors:
- Jean-Michel Dischler, Professor in Computer Science at the University of Strasbourg (dischler AT unistra.fr).
- Remi Allegre, Associate Professor in Computer Science at the University of Strasbourg (allegre AT unistra.fr);
- Jean-Michel Dischler, Professor in Computer Science at the University of Strasbourg (dischler AT unistra.fr).
- Remi Allegre, Associate Professor in Computer Science at the University of Strasbourg (allegre AT unistra.fr);
Starting date: September 2013.
Ending date: 3 years from the starting date.
Funding: The net salary will be about 1500 euros per month.
Location: Strasbourg area, France.
Application deadline: June 1, 2013
Description:
Current 3D acquisition technologies can deliver highly accurate data on the shape of real-world objects. Despite the recent advances in the field of geometry processing and 3D reconstruction, producing complete and defect-free multi-scale geometric models of objects that are as complex as pieces of artwork or buildings still cannot be achieved in a full automatic way, which hampers massive scanning campaigns. Standard 3d reconstruction techniques often require a large amount of user interaction to adjust parameters or to perform corrections, which essentially rely on some visual criteria. The recent advances of 2D/3D registration techniques open up new lines of research to bring solutions to the problem of automatically reconstructing complete and multi-scale accurate 3D models with few constraints on the acquisition process.
The goal of this thesis is to develop a technique that takes benefit from the information obtained from a set of photographs to evaluate and improve the reconstruction of 3D models from 3D scans. This technique will be based on a novel metric that allows to compare a calibrated set of photographs to a set of 3D scans. This technique will be studied in conjunction with a representation making it possible to visualize complex reconstructed models in real-time. The implementation will be achieved in C++ and it will be incorporated into the digitization software of the IGG group.
Current 3D acquisition technologies can deliver highly accurate data on the shape of real-world objects. Despite the recent advances in the field of geometry processing and 3D reconstruction, producing complete and defect-free multi-scale geometric models of objects that are as complex as pieces of artwork or buildings still cannot be achieved in a full automatic way, which hampers massive scanning campaigns. Standard 3d reconstruction techniques often require a large amount of user interaction to adjust parameters or to perform corrections, which essentially rely on some visual criteria. The recent advances of 2D/3D registration techniques open up new lines of research to bring solutions to the problem of automatically reconstructing complete and multi-scale accurate 3D models with few constraints on the acquisition process.
The goal of this thesis is to develop a technique that takes benefit from the information obtained from a set of photographs to evaluate and improve the reconstruction of 3D models from 3D scans. This technique will be based on a novel metric that allows to compare a calibrated set of photographs to a set of 3D scans. This technique will be studied in conjunction with a representation making it possible to visualize complex reconstructed models in real-time. The implementation will be achieved in C++ and it will be incorporated into the digitization software of the IGG group.
References:
[BWS07] A. Brunton, S. Wuhrer, C. Shu. Image-based Model Completion. In Proc. 3-D Digital Imaging and Modeling (3DIM), pages 305-311, 2007.
[CDG+12] M. Corsini, M. Dellepiane, F. Ganovelli, R. Gherardi, A. Fusiello, R. Scopigno. Fully Automatic Registration of Image Sets on Approximate Geometry. International Journal of Computer Vision, pages 1-21, August 10, 2012.
[DVS11] M. Dellepiane, A. Venturi, R. Scopigno. Image guided reconstruction of un-sampled data: A filling technique for cultural heritage models. International Journal of Computer Vision, 94(1):2-11, 2011.
[FP07] Y. Furukawa and J. Ponce. Accurate, Dense, and Robust Multi-View Stereopsis. IEEE Computer Society Conference on Computer Vision and Pattern Recognition, July 2007.
[PGC11] R. Pintus, E. Gobbetti, R. Combet. Fast and robust semi-automatic registration of photographs to 3D geometry. In Proc. International Symposium on Virtual Reality, Archaeology and Cultural Heritage (VAST), 2011.
[SSS06] N. Snavely, S. M. Seitz, R. Szeliski. Photo Tourism: Exploring Image Collections in 3D. ACM Transactions on Graphics (Proc. SIGGRAPH 2006), 2006.
[SY09] N. Salman and M. Yvinec. Surface Reconstruction from Multi-View Stereo. In Proc. Asian Conference on Computer Vision, Lecture Notes in Computer Science, 2009.
[VLP+12] H.-H. Vu, P. Labatut, J.-P. Pons, R. Keriven. High Accuracy and Visibility-Consistent Dense Multiview Stereo. IEEE Transactions on Pattern Analysis and Machine Intelligence, 34(5):889-901, 2012.
[BWS07] A. Brunton, S. Wuhrer, C. Shu. Image-based Model Completion. In Proc. 3-D Digital Imaging and Modeling (3DIM), pages 305-311, 2007.
[CDG+12] M. Corsini, M. Dellepiane, F. Ganovelli, R. Gherardi, A. Fusiello, R. Scopigno. Fully Automatic Registration of Image Sets on Approximate Geometry. International Journal of Computer Vision, pages 1-21, August 10, 2012.
[DVS11] M. Dellepiane, A. Venturi, R. Scopigno. Image guided reconstruction of un-sampled data: A filling technique for cultural heritage models. International Journal of Computer Vision, 94(1):2-11, 2011.
[FP07] Y. Furukawa and J. Ponce. Accurate, Dense, and Robust Multi-View Stereopsis. IEEE Computer Society Conference on Computer Vision and Pattern Recognition, July 2007.
[PGC11] R. Pintus, E. Gobbetti, R. Combet. Fast and robust semi-automatic registration of photographs to 3D geometry. In Proc. International Symposium on Virtual Reality, Archaeology and Cultural Heritage (VAST), 2011.
[SSS06] N. Snavely, S. M. Seitz, R. Szeliski. Photo Tourism: Exploring Image Collections in 3D. ACM Transactions on Graphics (Proc. SIGGRAPH 2006), 2006.
[SY09] N. Salman and M. Yvinec. Surface Reconstruction from Multi-View Stereo. In Proc. Asian Conference on Computer Vision, Lecture Notes in Computer Science, 2009.
[VLP+12] H.-H. Vu, P. Labatut, J.-P. Pons, R. Keriven. High Accuracy and Visibility-Consistent Dense Multiview Stereo. IEEE Transactions on Pattern Analysis and Machine Intelligence, 34(5):889-901, 2012.
Application process:
Candidates are invited to contact us as soon as possible via the two following email addresses: allegre AT unistra.fr and dischler AT unistra.fr. Please send us the following elements:
- a detailed CV;
- marks obtained during Licence and Master degree, or Engineering School degree;
- a one-page motivation letter;
- a recommendation letter from teachers.
Candidates are invited to contact us as soon as possible via the two following email addresses: allegre AT unistra.fr and dischler AT unistra.fr. Please send us the following elements:
- a detailed CV;
- marks obtained during Licence and Master degree, or Engineering School degree;
- a one-page motivation letter;
- a recommendation letter from teachers.
Qualification:
- Master degree in Computer Science;
- Basic skills in Computer Graphics and Geometric Modeling;
- Good programming and communication skills;
- Good level of English is mandatory.
- Master degree in Computer Science;
- Basic skills in Computer Graphics and Geometric Modeling;
- Good programming and communication skills;
- Good level of English is mandatory.
Application Deadline : 1 June 2013