3d landmark model discovery from a registered set of
play

3D Landmark Model Discovery from a Registered Set of Organic Shapes - PowerPoint PPT Presentation

Oct 30, 2023 •415 likes •704 views

3D Landmark Model Discovery from a Registered Set of Organic Shapes Clement Creusot, Nick Pears, Jim Austin Department of Computer science PCP, CVPRW, Providence (RI), June 2012 Plan What/Why How Results Conclusion References What/Why:

  1. 3D Landmark Model Discovery from a Registered Set of Organic Shapes Clement Creusot, Nick Pears, Jim Austin Department of Computer science PCP, CVPRW, Providence (RI), June 2012

  2. Plan What/Why How Results Conclusion References What/Why: Generalities and Problems How: Proposed method Results Conclusion Clement Creusot PCP, CVPRW, Providence (RI), June 2012, 2 / 20

  3. Generalities s What/Why Where is Wally? How Scene Results Waldo? Conclusion Charlie? References Walter? ウォ ー リ ー ? 威 利 ? . . . Clement Creusot PCP, CVPRW, Providence (RI), June 2012, 3 / 20

  4. Generalities s What/Why Where is Wally? How Output Scene Results Waldo? Conclusion Charlie? References Walter? ウォ ー リ ー ? 威 利 ? Detector . . . Clement Creusot PCP, CVPRW, Providence (RI), June 2012, 3 / 20

  5. Generalities s What/Why Where is Wally? How Output Scene Results Waldo? Conclusion Charlie? References Walter? ウォ ー リ ー ? 威 利 ? Detector . . . Model Clement Creusot PCP, CVPRW, Providence (RI), June 2012, 3 / 20

  6. Generalities s What/Why Where is Wally? How Output Scene Results Waldo? Conclusion Charlie? References Walter? ウォ ー リ ー ? 威 利 ? Detector . . . Model Model Discoverer Clement Creusot PCP, CVPRW, Providence (RI), June 2012, 3 / 20

  7. Generalities s What/Why Where is Wally? How Output Scene Results Waldo? Conclusion Charlie? References Walter? ウォ ー リ ー ? 威 利 ? Detector . . . Model ? Model Discoverer Clement Creusot PCP, CVPRW, Providence (RI), June 2012, 3 / 20

  8. Model Discovery for 3D Face Landmarking s What/Why How Output Scene Results Conclusion References Detector Model Clement Creusot PCP, CVPRW, Providence (RI), June 2012, 4 / 20

  9. Model Discovery for 3D Face Landmarking s What/Why How Output Scene Results Conclusion References Landmarker Model Clement Creusot PCP, CVPRW, Providence (RI), June 2012, 4 / 20

  10. Model Discovery for 3D Face Landmarking s What/Why How Output Scene Results Conclusion References Landmarker Model ? Model Discoverer Clement Creusot PCP, CVPRW, Providence (RI), June 2012, 4 / 20

  11. Why? - Gap in Research What/Why How Results Conclusion References [Amberg et al., 2007] [Creusot et al., 2011] [Gupta et al., 2007] [Romero-Huertas and Pears, 2008] [Szeptycki et al., 2009] [Zhao et al., 2011] Easy to label or explain to an operator Linked to 2D projections and plane symmetries Overall arbitrary Clement Creusot PCP, CVPRW, Providence (RI), June 2012, 5 / 20

  12. Nature of a model for a 3D-object class Sparse What/Why “Descriptive” How Results Conclusion Featural/Local information (nodes) References Structural/Global information (edges/hyperedges) Possible Local Features: Object Points Curves Surfaces Volumes Clement Creusot PCP, CVPRW, Providence (RI), June 2012, 6 / 20

  13. Nature of a model for a 3D-object class Sparse What/Why “Descriptive” How Results Conclusion Featural/Local information (nodes) References Structural/Global information (edges/hyperedges) Possible Local Features: Object Points Curves Surfaces Volumes Clement Creusot PCP, CVPRW, Providence (RI), June 2012, 6 / 20

  14. Organicly-shape objects What/Why How Results Conclusion More possible point-models than geometric shapes References Less intuition about what model is good Clement Creusot PCP, CVPRW, Providence (RI), June 2012, 7 / 20

  15. Example of 3D-objects point models Articulated Models: Non-Articulated Models: What/Why Articulations ??? How Results Extremities ??? Conclusion References [Shotton et al., 2011] [Bray et al., 2004] [Creusot et al., 2011] Clement Creusot PCP, CVPRW, Providence (RI), June 2012, 8 / 20

  16. Hypothesis What/Why Output Scene How Results Conclusion References Landmarker Model ? Model Discoverer Clement Creusot PCP, CVPRW, Providence (RI), June 2012, 9 / 20

  17. Hypothesis What/Why Output Scene “Probabilistic” How Results response map Conclusion available References One point per model Landmarker Model ? Model Discoverer Clement Creusot PCP, CVPRW, Providence (RI), June 2012, 9 / 20

  18. Our Approach What/Why How Results Use Detector and Neighborhood definition from Conclusion [Creusot et al., 2011] References 8 Local Descriptors Gaussian Distributions Linear Combination (LDA based) Test as many models as there are vertices in the template mesh ( ∼ 2000 ) Define two cost functions for each model: Saliency : Different from its neighborhood (good) Ubiquity : Ubiquitous over the face (bad) Clement Creusot PCP, CVPRW, Providence (RI), June 2012, 10 / 20

  19. Databases What/Why How Results Conclusion References FRGC (real) BFM (synthetic) (Coarse Correspondence) (Fine Correspondence) Clement Creusot PCP, CVPRW, Providence (RI), June 2012, 11 / 20

  20. Saliency Score per Vertex What/Why How Results 0.18 0.18 0.9 0.9 0.16 0.16 0.8 0.8 0.14 0.14 Conclusion 0.7 0.7 0.12 0.12 0.6 0.6 Density Density Density Density 0.1 0.1 0.5 0.5 References 0.08 0.08 0.4 0.4 0.06 0.06 0.3 0.3 0.04 0.04 0.2 0.2 0.02 0.02 0.1 0.1 0 0 0 0 0 0 0.2 0.2 0.4 0.4 0.6 0.6 0.8 0.8 1 1 0 0 0.2 0.2 0.4 0.4 0.6 0.6 0.8 0.8 1 1 Score Score Score Score 0.25 0.25 0.3 0.3 0.2 0.2 0.25 0.25 0.2 0.2 Density Density Density Density 0.15 0.15 0.15 0.15 0.1 0.1 0.1 0.1 0.05 0.05 0.05 0.05 0 0 0 0 0 0 0.2 0.2 0.4 0.4 0.6 0.6 0.8 0.8 1 1 0 0 0.2 0.2 0.4 0.4 0.6 0.6 0.8 0.8 1 1 0.962 0.981 1.00 Score Score 0.962 0.981 1.00 Score Score Clement Creusot PCP, CVPRW, Providence (RI), June 2012, 12 / 20

  21. Ubiquity Score per Vertex What/Why How Results Conclusion References 15.8 15.8 586. 586. 1.16e+03 1.16e+03 Clement Creusot PCP, CVPRW, Providence (RI), June 2012, 13 / 20

  22. Results Manual Automatic What/Why How Results Conclusion Initial References Symmetry Clement Creusot PCP, CVPRW, Providence (RI), June 2012, 14 / 20

  23. Saliency What/Why Similar How ID 8 3 1 2 5 6 7 Results U 658.37 215.52 287.70 67.452 252.50 393.68 324.40 Conclusion Automatic References ID 0 1 2 3 4 5 6 U 683.62 192.43 492.43 87.408 328.08 226.66 206.76 Manual Clement Creusot PCP, CVPRW, Providence (RI), June 2012, 15 / 20

  24. Saliency What/Why Different How 0 4 9 Results 325.51 561.31 416.18 U Conclusion Automatic References ID 7 8 9 701.64 954.16 444.54 U Manual Clement Creusot PCP, CVPRW, Providence (RI), June 2012, 16 / 20

  25. Problems What/Why How Different answers depending on the registration Results method: Conclusion Fine registration on clean data (BFM) References Coarse registration on unclean data (FRGC) Fine registration on unclean data (???) Needed Optimization method → Depends on the detector used (and its parameters) How to include structural information in the model discovery? How to project a newly discovered model to unseen training data? (again a registration problem) Clement Creusot PCP, CVPRW, Providence (RI), June 2012, 17 / 20

  26. Conclusion What/Why Good: How Optimize model for a detector Results Validate most human-chosen landmarks Conclusion Give quantifiable measure of landmark quality References Bad: Only non-articulated objects for now Requires a large set of finely-registered objects (Do you have one to share?) Questions to you: How do you learn a model structure in your application domain? Are there applications where you think this might help? Brain teaser: How do you extend the idea to multi-dimensional features (curves, area, volumes)? Clement Creusot PCP, CVPRW, Providence (RI), June 2012, 18 / 20

  27. References I What/Why Amberg, B., Romdhani, S., and Vetter, T. (2007). Optimal step nonrigid icp algorithms for surface registration. How In IEEE Int Conf. CVPR . Results Bray, M., Koller-Meier, E., Mueller, P., Gool, L. V., and Schraudolph, N. N. Conclusion (2004). 3d hand tracking by rapid stochastic gradient descent using a skinning model. References In Chambers, A. and Hilton, A., editors, 1st European Conference on Visual Media Production (CVMP) , pages 59–68. IEE. Creusot, C., Pears, N., and Austin, J. (2011). Automatic keypoint detection on 3d faces using a dictionary of local shapes. In 3DIMPVT , pages 204–211. Gupta, S., Markey, M. K., Aggarwal, J., and Bovik, A. C. (2007). Three dimensional face recognition based on geodesic and euclidean distances. In IS&T/SPIE Symp. on Electronic Imaging: Vision Geometry XV . Romero-Huertas, M. and Pears, N. (2008). 3d facial landmark localisation by matching simple descriptors. In IEEE Int. Conf. BTAS , pages 1–6. Shotton, J., Fitzgibbon, A., Cook, M., Sharp, T., Finocchio, M., Moore, R., Kipman, A., and Blake, A. (2011). Real-time human pose recognition in parts from single depth images. In Computer Vision and Pattern Recognition (CVPR), 2011 IEEE Conference on , pages 1297 –1304. Clement Creusot PCP, CVPRW, Providence (RI), June 2012, 19 / 20

Download Presentation
Download Policy: The content available on the website is offered to you 'AS IS' for your personal information and use only. It cannot be commercialized, licensed, or distributed on other websites without prior consent from the author. To download a presentation, simply click this link. If you encounter any difficulties during the download process, it's possible that the publisher has removed the file from their server.

Recommend

Landmark Landmark-based routing based routing Landmark Landmark-based routing based routing

Landmark Landmark-based routing based routing Landmark Landmark-based routing based routing

Landmark Landmark-based routing based routing Landmark Landmark-based routing based routing [Kleinberg04] J. Kleinberg, A. Slivkins, T. Wexler. Triangulation and Embedding using Small Sets of Beacons. Proc. 45th IEEE Symposium on Foundations of

450 views • 32 slides
Model Architectures and Training Techniques for High-Precision Landmark Localization Sina Honari

Model Architectures and Training Techniques for High-Precision Landmark Localization Sina Honari

Model Architectures and Training Techniques for High-Precision Landmark Localization Sina Honari Pavlo Molchanov Jason Yosinski Stephen Tyree Pascal Vincent Jan Kautz Christopher Pal KEYPOINT DETECTION / LANDMARK LOCALIZATION The problem

605 views • 37 slides
CS 557 Landmark Routing The Landmark Hierarchy: A New Hierarchy For Routing in Very Large

CS 557 Landmark Routing The Landmark Hierarchy: A New Hierarchy For Routing in Very Large

CS 557 Landmark Routing The Landmark Hierarchy: A New Hierarchy For Routing in Very Large Networks Paul Tsuchiya, 1988 Spring 2013 Landmark Routing Objective: Reduce the routing table size without increasing path length by a large

458 views • 15 slides
The More the Merrier?! Evaluating the Effect of Landmark Extraction Algorithms on Landmark-Based

The More the Merrier?! Evaluating the Effect of Landmark Extraction Algorithms on Landmark-Based

The More the Merrier?! Evaluating the Effect of Landmark Extraction Algorithms on Landmark-Based Goal Recognition Kin Max Piamolini Gusm ao Ramon Fraga Pereira Felipe Meneguzzi Pontifical Catholic University of Rio Grande do Sul - Brazil

436 views • 21 slides
Topic 1: Landmark Designation Study more flexible alternatives to landmark districts Add

Topic 1: Landmark Designation Study more flexible alternatives to landmark districts Add

April 4, 2018 Historic Preservation Code Review Karen McWilliams, Historic Preservation Manager Recommending Topic 1: Landmark Designation Study more flexible alternatives to landmark districts Add time to multiple property designations:

391 views • 9 slides
RNA Search and Whirlwind tour of ncRNA search & discovery Motif Discovery RNA motif

RNA Search and Whirlwind tour of ncRNA search & discovery Motif Discovery RNA motif

Outline RNA Search and Whirlwind tour of ncRNA search & discovery Motif Discovery RNA motif description (Covariance Model Review) Algorithms for searching Rigorous & heuristic filtering Motif discovery Lecture

809 views • 27 slides
Higgs discovery and BSM Mihoko M. Nojiri 12 11 12 Higgs discovery at the

Higgs discovery and BSM Mihoko M. Nojiri 12 11 12 Higgs discovery at the

Higgs discovery and BSM Mihoko M. Nojiri 12 11 12 Higgs discovery at the LHC Higgs boson: The Last missing particle of the SM particles Probably starting point of the Beyond the stard model why we think

437 views • 31 slides
Landmark Map L11: Landmark Mapping Locations and uncertainties of n landmarks, with respect

Landmark Map L11: Landmark Mapping Locations and uncertainties of n landmarks, with respect

Landmark Map L11: Landmark Mapping Locations and uncertainties of n landmarks, with respect to a specific frame of reference. World frame: fixed origin point Robot frame: origin at the robot CS 344R/393R: Robotics Benjamin

363 views • 5 slides
Landmark-based Meta Best-First Search Bachelor Thesis Presentation By Samuel Hugger STRIPS

Landmark-based Meta Best-First Search Bachelor Thesis Presentation By Samuel Hugger STRIPS

Landmark-based Meta Best-First Search Bachelor Thesis Presentation By Samuel Hugger STRIPS Planning Model States Set of atoms True or False Actions Preconditions Add-efgects Delete-efgects STRIPS Planning Model

376 views • 26 slides
ON TRACK FOR DISCOVERY http://dsmsyndicate.ca/ February 2018 1 PROJECT GENERATOR MODEL The

ON TRACK FOR DISCOVERY http://dsmsyndicate.ca/ February 2018 1 PROJECT GENERATOR MODEL The

ON TRACK FOR DISCOVERY http://dsmsyndicate.ca/ February 2018 1 PROJECT GENERATOR MODEL The DSM syndicate is a project generator focused on original discovery resulting from glacial and snowpack recession. DSM management , exploration,

559 views • 15 slides
ON TRACK FOR DISCOVERY http://dsmsyndicate.ca/ February 2018 1 PROJECT GENERATOR MODEL The

ON TRACK FOR DISCOVERY http://dsmsyndicate.ca/ February 2018 1 PROJECT GENERATOR MODEL The

ON TRACK FOR DISCOVERY http://dsmsyndicate.ca/ February 2018 1 PROJECT GENERATOR MODEL The DSM syndicate is a project generator focused on original discovery resulting from glacial and snowpack recession. DSM management , exploration,

194 views • 16 slides
THE LANDMARK NICOSIA HOTEL POSITIONING The Landmark Nicosia, a 5* Hotel in Cyprus, is the

THE LANDMARK NICOSIA HOTEL POSITIONING The Landmark Nicosia, a 5* Hotel in Cyprus, is the

THE LANDMARK NICOSIA HOTEL POSITIONING The Landmark Nicosia, a 5* Hotel in Cyprus, is the ultimate choice for upper scale clientele, governmental business and all VIP visits. Your first choice both for business and leisure travel. CYPRUS

1.5k views • 12 slides
Supplementary Presentation Outscapes Horticultural Solutions & Landmark A Proven Relationship

Supplementary Presentation Outscapes Horticultural Solutions & Landmark A Proven Relationship

Request for Tender for Grounds Maintenance Services Landmark - Units Plan 2413 3-9 Bowen Drive and 41-47 Blackall Street Barton ACT 2600 Supplementary Presentation Outscapes Horticultural Solutions & Landmark A Proven Relationship

535 views • 24 slides
On the fate of the Standard Model in view of the Higgs discovery Gino Isidori [ INFN, Frascati

On the fate of the Standard Model in view of the Higgs discovery Gino Isidori [ INFN, Frascati

Warsaw, March 2013 G. Isidori On the fate of the Standard Model in view of the Higgs discovery On the fate of the Standard Model in view of the Higgs discovery Gino Isidori [ INFN, Frascati & CERN ] Introduction The Higgs potential

638 views • 53 slides
CAPT Xie Yanlu Beijing Language and Culture University Outline English landmark Methods

CAPT Xie Yanlu Beijing Language and Culture University Outline English landmark Methods

Landmark in Chinese CAPT Xie Yanlu Beijing Language and Culture University Outline English landmark Methods to select Chinese landmark Experiments in Chinese CAPT Discussion 2016/2/11 3 Objective in using computer aided

708 views • 42 slides
Opportunities and Challenges in Multi-Model, Multi- Dimensional Image Analysis for Drug Discovery

Opportunities and Challenges in Multi-Model, Multi- Dimensional Image Analysis for Drug Discovery

Opportunities and Challenges in Multi-Model, Multi- Dimensional Image Analysis for Drug Discovery Dr Yinhai Wang (Senior Scientist, Image Analytics) 9 th March 2016 Quantitative Biology Department, Discovery Sciences the Quantitative Biology

601 views • 29 slides
Classification and statistical machine learning Sylvain Arlot http://www.di.ens.fr/~arlot/ 1 Cnrs

Classification and statistical machine learning Sylvain Arlot http://www.di.ens.fr/~arlot/ 1 Cnrs

Introduction Goals Overfitting Examples Key issues Conclusion Classification and statistical machine learning Sylvain Arlot http://www.di.ens.fr/~arlot/ 1 Cnrs 2 erieure (Paris), DI/ENS , Ecole Normale Sup Equipe Sierra CEMRACS

1.57k views • 125 slides
CSC 411: Lecture 06: Decision Trees Class based on Raquel Urtasun & Rich Zemels lectures

CSC 411: Lecture 06: Decision Trees Class based on Raquel Urtasun & Rich Zemels lectures

CSC 411: Lecture 06: Decision Trees Class based on Raquel Urtasun & Rich Zemels lectures Sanja Fidler University of Toronto Jan 26, 2016 Urtasun, Zemel, Fidler (UofT) CSC 411: 06-Decision Trees Jan 26, 2016 1 / 39 Today Decision

542 views • 39 slides
of human actions Ivan Laptev ivan.laptev@inria.fr WILLOW, INRIA/ENS/CNRS, Paris Objects:

of human actions Ivan Laptev ivan.laptev@inria.fr WILLOW, INRIA/ENS/CNRS, Paris Objects:

ENS/INRIA CVML Summer School 45 rue dUlm , Paris July 26, 2013 Modeling and visual recognition of human actions Ivan Laptev ivan.laptev@inria.fr WILLOW, INRIA/ENS/CNRS, Paris Objects: Actions: cars, glasses, drinking, running, people,

1.61k views • 99 slides
Trans ansfor orming ming Tec echnic hnical al As Assis sistan tance: ce: Using Using

Trans ansfor orming ming Tec echnic hnical al As Assis sistan tance: ce: Using Using

Trans ansfor orming ming Tec echnic hnical al As Assis sistan tance: ce: Using Using Evidence Evidence to Enric to Enrich h the Experience the Experience Center for I Imp mproving ing Research Evide idence (CIRE (CI RE) )

936 views • 50 slides
Augusta Maine Short Handed Firefighting National Volunteer Fire Council Concord North Carolina

Augusta Maine Short Handed Firefighting National Volunteer Fire Council Concord North Carolina

Pine Point Fire Training Augusta Maine Short Handed Firefighting National Volunteer Fire Council Concord North Carolina Thank You National Volunteer Fire Council Pine Point Fire Training Augusta Greenville Maine

2.19k views • 153 slides
Ignition from a Fire Perimeter in a WRF Wildland Fire Model Volodymyr Kondratenko y y Based on

Ignition from a Fire Perimeter in a WRF Wildland Fire Model Volodymyr Kondratenko y y Based on

Department of Mathematical and Statistical Sciences Department of Mathematical and Statistical Sciences University of Colorado Denver University of Colorado Denver y Ignition from a Fire Perimeter in a WRF Wildland Fire Model Volodymyr

432 views • 14 slides
Coupled fire-atmosphere-fuel moisture-smoke online modeling with WRF-SFIRE Jan Mandel,

Coupled fire-atmosphere-fuel moisture-smoke online modeling with WRF-SFIRE Jan Mandel,

Coupled fire-atmosphere-fuel moisture-smoke online modeling with WRF-SFIRE Jan Mandel, University of Colorado Denver Adam K. Kochanski, University of Utah Sher Schranz, NOAA/CIRA Martin Vejmelka, AVAST Supported by NASA grant NNX13AH59G and

613 views • 30 slides
Nation National al Fire Fire Protection Protection Assoc Association iation (NFPA) Standard

Nation National al Fire Fire Protection Protection Assoc Association iation (NFPA) Standard

Bri Briefing efing on on Nation National al Fire Fire Protection Protection Assoc Association iation (NFPA) Standard (NFPA) Standard 805 Fire Protection 805 Fire Protection Mark Satorius Executive Director for Operations, OEDO June

423 views • 16 slides

More recommend