Explore

Paper Presentation
Paper Presentation

An important aspect of research is to read papers that have been published in the field to stay updated with the state of the art as well as to explore new or different ideas.

In this assignment, you will have the opportunity to read, critically analyze, and present a paper on geometric computer vision.

Remember that, the publication process is dependent on peer-review, where fellow researchers review the papers submitted to a conference/journal, and the decision on the acceptance/rejection is based on the reviews. Thus it is crucial for the field to have reviewers that evaluate the papers critically and thoroughly.

Objective

The goal of this assignment is to:

explore and become more familiar with the state of the art in geometric computer vision

gain experience in reading and reviewing papers.

In the process, you will also be able to work on your presentation and teamwork skills.

Assignment

You will work in groups of 2-3 students to present a recent paper in the area of geometric computer vision. The idea is to read the selected paper with a critical lens: to understand the proposed method and its context in the literature, analyze its strengths and weaknesses, and to come up with further questions or even ideas to improve it. You can also try the implementation of the method (the code may be publicly available or you can request the authors for it) to test it and thereby understand it better.

I’ve provided a list of suggested papers at the end of this doc, but you can also select a paper not on this list, provided I approve it (see the “Paper proposal” section in the “To-do” below). Broadly, the criteria to select a paper is:

Published at a top conference (CVPR/ICCV/ECCV) or journal (IJCV, PAMI) in computer vision

Relatively recent (after ~2015)

Preferably on a geometric/algebraic approach

If you are really keen on presenting a paper that does not meet the above criteria, you can still submit it in your proposal and I will consider it.

To-do

The assignment will have three phases:

Paper proposal

Discussion session

Paper presentation

Paper proposal

Each group will submit a proposal that contains a list of 3 preferred papers that the group would like to work on, along with a brief comment on the reason for selecting each paper.

The list of papers should be be sorted in order of preference (1→most preferred). If multiple groups have the same first preference, it will be assigned to the group that submitted the proposal first.

Proposal format: PDF file containing the names of the group members and the ordered list of papers, along with a brief comment on the reasons for selecting each paper. For each paper, include details such as author names, conference/journal, year of publication etc. (you can get this easily from Google scholar, as done for the “Suggested papers” list below). The reason for selecting each paper should be short and clear - 1 or 2 lines are sufficient. A template in Latex for the proposal is available on the VP channel in MS Teams.

Send your proposal by email to:

Devesh.Adlakha@u-bourgogne.fr⁠

⁠

david.fofi@u-bourgogne.fr⁠

⁠

Deadline: Tuesday, 30 March 23h59

Discussion session

After your paper proposals have been submitted (and approved), we will have a session where you can work on your paper presentations and ask any questions related to the assignment or your papers.

During this session, each group will discuss their progress with me, including:

Work done so far

Work remaining

Estimated timeline

Project management (tools used, division of work between group members, etc.)

I will assign a grade (as project management) based on this discussion, which will factor in to the final grade for the paper presentation.

Discussion session: 27 April 10h-12h

Paper presentation

The paper presentations will be in front of a jury composed of me, David Fofi, and possibly other researchers from the VIBOT team.

Evaluation criteria: includes demonstrating an understanding of the selected paper, its context in the literature, critical analysis of the work, and discussion and ideas for extensions or improvements.

Format: each group will have 15 minutes maximum for the presentation followed by ~10 minutes for Q&A. We will interrupt and stop the presentation after 15 minutes (this happens sometimes at conference sessions as well!).

Presentation date and location: Monday 31 May 14h-17h, room TBD

Schedule: I highly encourage you to attend all the presentations, and, if you have time, to read or at least skim through the papers presented by the other groups. This will enable you to learn more about the different papers presented by the other groups and have a wider understanding of geometric computer vision problems and methods. However, you do not have to attend the other presentations and you will be free to arrive in time for your own group’s presentation and leave after it. To avoid any delays, I suggest that you plan to arrive ahead of time by at least two presentations. For example, if your group is presenting fifth, ensure that you are there by the time of the third presentation (each presentation is of maximum 15 minutes, so you can estimate the time when you should be there).

Each group will have a couple of minutes to set up. Make sure that you have everything you need - adapters to connect to the projector, laser pointer, etc. I suggest that you test your set up in advance in the presentation room (room to be decided, probably will be Amphi 8).

The following is the order of the presentations (which I determined through a random wheel spin, similar to how we did it for the OpenCV project presentations):

Lateef and Fatai

S. Agrawal, A. Pahuja and S. Lucey, "High Accuracy Face Geometry Capture using a Smartphone Video," 2020 IEEE Winter Conference on Applications of Computer Vision (WACV), 2020, pp. 81-90, doi: 10.1109/WACV45572.2020.9093455.

https://arxiv.org/pdf/2003.08583.pdf⁠

⁠

Walid, Fred, and Etinosa

Barath, Daniel. "Five-point fundamental matrix estimation for uncalibrated cameras." In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 235-243. 2018.

https://openaccess.thecvf.com/content_cvpr_2018/papers/Barath_Five-Point_Fundamental_Matrix_CVPR_2018_paper.pdf⁠

⁠

Quentin, Dylan, and Camille

Peng, Songyou, and Peter Sturm. "Calibration Wizard: A guidance system for camera calibration based on modelling geometric and corner uncertainty." In Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 1497-1505. 2019.

https://openaccess.thecvf.com/content_ICCV_2019/papers/Peng_Calibration_Wizard_A_Guidance_System_for_Camera_Calibration_Based_on_ICCV_2019_paper.pdf⁠

⁠

Yagmur, Divine, and Solomon

Magerand, Ludovic, and Alessio Del Bue. "Practical projective structure from motion (p2sfm)." In Proceedings of the IEEE International Conference on Computer Vision, pp. 39-47. 2017.

https://openaccess.thecvf.com/content_ICCV_2017/papers/Magerand_Practical_Projective_Structure_ICCV_2017_paper.pdf⁠

⁠

Arul, Sofia, and Moiz

Cohen, Andrea, Johannes L. Schönberger, Pablo Speciale, Torsten Sattler, Jan-Michael Frahm, and Marc Pollefeys. "Indoor-outdoor 3d reconstruction alignment." In European Conference on Computer Vision, pp. 285-300. Springer, Cham, 2016.

https://www.researchgate.net/profile/Johannes-Schoenberger/publication/305655690_Indoor-Outdoor_3D_Reconstruction_Alignment/links/5798549e08aec89db7b9b6d6/Indoor-Outdoor-3D-Reconstruction-Alignment.pdf⁠

⁠

Suggested papers

I will add more papers to this list, so you can check this page regularly before selecting a paper.

Kasten, Yoni, Amnon Geifman, Meirav Galun, and Ronen Basri. "Algebraic characterization of essential matrices and their averaging in multiview settings." In Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 5895-5903. 2019.

https://openaccess.thecvf.com/content_ICCV_2019/papers/Kasten_Algebraic_Characterization_of_Essential_Matrices_and_Their_Averaging_in_Multiview_ICCV_2019_paper.pdf⁠

⁠

Kasten, Yoni, Amnon Geifman, Meirav Galun, and Ronen Basri. "Gpsfm: Global projective sfm using algebraic constraints on multi-view fundamental matrices." In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 3264-3272. 2019.

https://openaccess.thecvf.com/content_CVPR_2019/papers/Kasten_GPSfM_Global_Projective_SFM_Using_Algebraic_Constraints_on_Multi-View_Fundamental_CVPR_2019_paper.pdf⁠

⁠

Magerand, Ludovic, and Alessio Del Bue. "Practical projective structure from motion (p2sfm)." In Proceedings of the IEEE International Conference on Computer Vision, pp. 39-47. 2017.

https://openaccess.thecvf.com/content_ICCV_2017/papers/Magerand_Practical_Projective_Structure_ICCV_2017_paper.pdf⁠

⁠

Geppert, Marcel, Viktor Larsson, Pablo Speciale, Johannes L. Schönberger, and Marc Pollefeys. "Privacy preserving structure-from-motion." In European Conference on Computer Vision, pp. 333-350. Springer, Cham, 2020.

https://cvg.ethz.ch/research/privacy-preserving-sfm/paper/Geppert2020ECCV.pdf⁠

⁠

Shi, Yunpeng, and Gilad Lerman. "Estimation of camera locations in highly corrupted scenarios: All about that base, no shape trouble." In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 2868-2876. 2018.

https://openaccess.thecvf.com/content_cvpr_2018/papers/Shi_Estimation_of_Camera_CVPR_2018_paper.pdf⁠

⁠

Speciale, Pablo, Johannes L. Schonberger, Sing Bing Kang, Sudipta N. Sinha, and Marc Pollefeys. "Privacy preserving image-based localization." In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 5493-5503. 2019.

https://openaccess.thecvf.com/content_CVPR_2019/papers/Speciale_Privacy_Preserving_Image-Based_Localization_CVPR_2019_paper.pdf⁠

⁠

Schops, Thomas, Viktor Larsson, Marc Pollefeys, and Torsten Sattler. "Why having 10,000 parameters in your camera model is better than twelve." In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 2535-2544. 2020.

https://openaccess.thecvf.com/content_CVPR_2020/papers/Schops_Why_Having_10000_Parameters_in_Your_Camera_Model_Is_Better_CVPR_2020_paper.pdf⁠

⁠

https://openaccess.thecvf.com/content_ICCV_2019/papers/Peng_Calibration_Wizard_A_Guidance_System_for_Camera_Calibration_Based_on_ICCV_2019_paper.pdf⁠

⁠

Cavalli, Luca, Viktor Larsson, Martin Ralf Oswald, Torsten Sattler, and Marc Pollefeys. "Handcrafted Outlier Detection Revisited." In European Conference on Computer Vision, pp. 770-787. Springer, Cham, 2020.

https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123640766.pdf⁠

⁠

Arrigoni, Federica, Luca Magri, and Tomas Pajdla. "On the Usage of the Trifocal Tensor in Motion Segmentation." Computer Vision—ECCV (2020).

http://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123650511.pdf⁠

⁠

Albl, Cenek, Zuzana Kukelova, Viktor Larsson, Michal Polic, Tomas Pajdla, and Konrad Schindler. "From two rolling shutters to one global shutter." In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 2505-2513. 2020.

https://openaccess.thecvf.com/content_CVPR_2020/papers/Albl_From_Two_Rolling_Shutters_to_One_Global_Shutter_CVPR_2020_paper.pdf⁠

⁠

Martyushev, Evgeniy. "Self-calibration of cameras with Euclidean image plane in case of two views and known relative rotation angle." In Proceedings of the European Conference on Computer Vision (ECCV), pp. 415-429. 2018.

http://openaccess.thecvf.com/content_ECCV_2018/papers/Martyushev_Self-Calibration_of_Cameras_ECCV_2018_paper.pdf⁠

⁠

Maset, Eleonora, Federica Arrigoni, and Andrea Fusiello. "Practical and efficient multi-view matching." In Proceedings of the IEEE International Conference on Computer Vision, pp. 4568-4576. 2017.

https://openaccess.thecvf.com/content_ICCV_2017/papers/Maset_Practical_and_Efficient_ICCV_2017_paper.pdf⁠

⁠

Arrigoni, Federica, Beatrice Rossi, and Andrea Fusiello. "Global registration of 3D point sets via LRS decomposition." In European Conference on Computer Vision, pp. 489-504. Springer, Cham, 2016.

http://www.diegm.uniud.it/fusiello/papers/eccv16.pdf⁠

⁠

Magri, Luca, and Andrea Fusiello. "Fitting multiple heterogeneous models by multi-class cascaded t-linkage." In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 7460-7468. 2019.

http://openaccess.thecvf.com/content_CVPR_2019/papers/Magri_Fitting_Multiple_Heterogeneous_Models_by_Multi-Class_Cascaded_T-Linkage_CVPR_2019_paper.pdf⁠

⁠

Camposeco, Federico, Andrea Cohen, Marc Pollefeys, and Torsten Sattler. "Hybrid camera pose estimation." In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 136-144. 2018.

https://openaccess.thecvf.com/content_cvpr_2018/papers/Camposeco_Hybrid_Camera_Pose_CVPR_2018_paper.pdf⁠

⁠

Cohen, Andrea, Torsten Sattler, and Marc Pollefeys. "Merging the unmatchable: Stitching visually disconnected sfm models." In Proceedings of the IEEE International Conference on Computer Vision, pp. 2129-2137. 2015.

https://openaccess.thecvf.com/content_iccv_2015/papers/Cohen_Merging_the_Unmatchable_ICCV_2015_paper.pdf⁠

⁠

Barath, Daniel. "Five-point fundamental matrix estimation for uncalibrated cameras." In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 235-243. 2018.

https://openaccess.thecvf.com/content_cvpr_2018/papers/Barath_Five-Point_Fundamental_Matrix_CVPR_2018_paper.pdf⁠

⁠

Barath, Daniel, and Jiri Matas. "Progressive-X: Efficient, anytime, multi-model fitting algorithm." In Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 3780-3788. 2019.

https://openaccess.thecvf.com/content_ICCV_2019/papers/Barath_Progressive-X_Efficient_Anytime_Multi-Model_Fitting_Algorithm_ICCV_2019_paper.pdf⁠

⁠

Ding, Yaqing, Daniel Barath, Jian Yang, Hui Kong, and Zuzana Kukelova. "Globally Optimal Relative Pose Estimation with Gravity Prior." CVPR 2021.

https://arxiv.org/pdf/2012.00458.pdf⁠

⁠

Polic, Michal, Stanislav Steidl, Cenek Albl, Zuzana Kukelova, and Tomas Pajdla. "Uncertainty Based Camera Model Selection." In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 5991-6000. 2020.

https://openaccess.thecvf.com/content_CVPR_2020/papers/Polic_Uncertainty_Based_Camera_Model_Selection_CVPR_2020_paper.pdf⁠

⁠

Polic, Michal, Wolfgang Forstner, and Tomas Pajdla. "Fast and accurate camera covariance computation for large 3d reconstruction." In Proceedings of the European Conference on Computer Vision (ECCV), pp. 679-694. 2018.

https://openaccess.thecvf.com/content_ECCV_2018/papers/Michal_Polic_Fast_and_Precise_ECCV_2018_paper.pdf⁠

⁠

Albl, Cenek, Akihiro Sugimoto, and Tomas Pajdla. "Degeneracies in rolling shutter sfm." In European Conference on Computer Vision, pp. 36-51. Springer, Cham, 2016.

https://people.ciirc.cvut.cz/~pajdla/papers/Albl-ECCV-2016.pdf⁠

⁠

Crocco, Marco, Cosimo Rubino, and Alessio Del Bue. "Structure from motion with objects." In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 4141-4149. 2016.

https://www.cv-foundation.org/openaccess/content_cvpr_2016/papers/Crocco_Structure_From_Motion_CVPR_2016_paper.pdf⁠

⁠

Lee, Seong Hun, and Javier Civera. "Closed-form optimal two-view triangulation based on angular errors." In Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 2681-2689. 2019.

https://openaccess.thecvf.com/content_ICCV_2019/papers/Lee_Closed-Form_Optimal_Two-View_Triangulation_Based_on_Angular_Errors_ICCV_2019_paper.pdf⁠

⁠

Lee, Seong Hun, and Javier Civera. "Rotation-Only Bundle Adjustment." CVPR 2021.

https://arxiv.org/pdf/2011.11724.pdf⁠

⁠

Ben-Artzi, Gil. "Separable Four Points Fundamental Matrix." In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pp. 188-196. 2021.

https://openaccess.thecvf.com/content/WACV2021/papers/Ben-Artzi_Separable_Four_Points_Fundamental_Matrix_WACV_2021_paper.pdf⁠

⁠

Want to print your doc?
This is not the way.

Try clicking the ⋯ next to your doc name or using a keyboard shortcut (

CtrlP

) instead.

Paper PresentationPaper Presentation

Objective

Assignment

To-do

Paper proposal

Discussion session

Paper presentation

Suggested papers

Paper Presentation
Paper Presentation