SlideShare a Scribd company logo
Real-time tracking
Examples
Which norm should we use?
Real-time Tracking with Features
Petter Strandmark
Lund University
February 28, 2009
Petter Strandmark Real-time Tracking with Features
Real-time tracking
Examples
Which norm should we use?
Outline
1 Real-time tracking
SIFT (review)
Estimating transformations
RANSAC (review)
SURF
2 Examples
Successful examples
Failures
Multiple objects
3 Which norm should we use?
Line fitting
Panorama stitching
Petter Strandmark Real-time Tracking with Features
Real-time tracking
Examples
Which norm should we use?
SIFT (review)
Estimating transformations
RANSAC (review)
SURF
SIFT
Petter Strandmark Real-time Tracking with Features
Real-time tracking
Examples
Which norm should we use?
SIFT (review)
Estimating transformations
RANSAC (review)
SURF
Estimating transformations
Petter Strandmark Real-time Tracking with Features
Real-time tracking
Examples
Which norm should we use?
SIFT (review)
Estimating transformations
RANSAC (review)
SURF
RANSAC
RAndom SAmple Consensus
Least−squares Least−squares RANSAC
Petter Strandmark Real-time Tracking with Features
Real-time tracking
Examples
Which norm should we use?
SIFT (review)
Estimating transformations
RANSAC (review)
SURF
SURF
Gaussian Box filter Gaussian Box filter
Approximate Gaussians with box filters.
Petter Strandmark Real-time Tracking with Features
Real-time tracking
Examples
Which norm should we use?
Successful examples
Failures
Multiple objects
“David”
Tracking example
Petter Strandmark Real-time Tracking with Features
Real-time tracking
Examples
Which norm should we use?
Successful examples
Failures
Multiple objects
Tracking vs. recognition
Recognition
Tracking
Petter Strandmark Real-time Tracking with Features
Real-time tracking
Examples
Which norm should we use?
Successful examples
Failures
Multiple objects
“All-Bran”
Result
Petter Strandmark Real-time Tracking with Features
Real-time tracking
Examples
Which norm should we use?
Successful examples
Failures
Multiple objects
“Pedestrian”
Result
Petter Strandmark Real-time Tracking with Features
Real-time tracking
Examples
Which norm should we use?
Successful examples
Failures
Multiple objects
Rotation
Is handled well
Petter Strandmark Real-time Tracking with Features
Real-time tracking
Examples
Which norm should we use?
Successful examples
Failures
Multiple objects
Illumination
SIFT descriptors are invariant to illumination
Petter Strandmark Real-time Tracking with Features
Real-time tracking
Examples
Which norm should we use?
Successful examples
Failures
Multiple objects
Occlusions
Only a subset of the SIFT descriptors need to be found
Petter Strandmark Real-time Tracking with Features
Real-time tracking
Examples
Which norm should we use?
Successful examples
Failures
Multiple objects
Untextured Objects
Unable to extract SIFT features
Petter Strandmark Real-time Tracking with Features
Real-time tracking
Examples
Which norm should we use?
Successful examples
Failures
Multiple objects
Multiple objects
Without any modification
Petter Strandmark Real-time Tracking with Features
Real-time tracking
Examples
Which norm should we use?
Successful examples
Failures
Multiple objects
Multiple objects
Additional considerations
Petter Strandmark Real-time Tracking with Features
Real-time tracking
Examples
Which norm should we use?
Successful examples
Failures
Multiple objects
Multiple objects
Only add points within a single frame when updating
Petter Strandmark Real-time Tracking with Features
Real-time tracking
Examples
Which norm should we use?
Line fitting
Panorama stitching
Which norm should we use?
Petter Strandmark Real-time Tracking with Features
Real-time tracking
Examples
Which norm should we use?
Line fitting
Panorama stitching
Which norm should we use?
|| · ||∞
Petter Strandmark Real-time Tracking with Features
Real-time tracking
Examples
Which norm should we use?
Line fitting
Panorama stitching
Which norm should we use?
|| · ||2
Petter Strandmark Real-time Tracking with Features
Real-time tracking
Examples
Which norm should we use?
Line fitting
Panorama stitching
Which norm should we use?
|| · ||1
Petter Strandmark Real-time Tracking with Features
Real-time tracking
Examples
Which norm should we use?
Line fitting
Panorama stitching
Which norm should we use?
Petter Strandmark Real-time Tracking with Features
Real-time tracking
Examples
Which norm should we use?
Line fitting
Panorama stitching
Which norm should we use?
|| · ||∞
Petter Strandmark Real-time Tracking with Features
Real-time tracking
Examples
Which norm should we use?
Line fitting
Panorama stitching
Which norm should we use?
|| · ||2
Petter Strandmark Real-time Tracking with Features
Real-time tracking
Examples
Which norm should we use?
Line fitting
Panorama stitching
Which norm should we use?
|| · ||1
Petter Strandmark Real-time Tracking with Features
Real-time tracking
Examples
Which norm should we use?
Line fitting
Panorama stitching
Resources
Course home page
SURFmex :
http://users.student.lth.se/p04pst/surfmex.html
YALMIP optimization package :
http://control.ee.ethz.ch/~joloef/wiki/pmwiki.php
Petter Strandmark Real-time Tracking with Features

More Related Content

Viewers also liked

Sol amortiguadoras
Sol amortiguadorasSol amortiguadoras
Sol amortiguadorasUAM-I
 
Alexandra Velnidou on Service Design
Alexandra Velnidou on Service DesignAlexandra Velnidou on Service Design
Alexandra Velnidou on Service Design
Alexandra Velnidou
 
Tugas kkpi
Tugas kkpiTugas kkpi
Tugas kkpi
Ma'ul Maulana
 
A Personal Testimonial From a Morgan Rothschild Academy Parent
A Personal Testimonial From a Morgan Rothschild Academy ParentA Personal Testimonial From a Morgan Rothschild Academy Parent
A Personal Testimonial From a Morgan Rothschild Academy Parent
Morgan Rothschild Academy
 
Cuisine then/now
Cuisine then/nowCuisine then/now
Cuisine then/now
nathalieruvalcaba
 
All budget-chart-book-2012
All budget-chart-book-2012All budget-chart-book-2012
All budget-chart-book-2012
Global Business & Co.
 
Oil, gas+power customer presentation email version
Oil, gas+power customer presentation   email versionOil, gas+power customer presentation   email version
Oil, gas+power customer presentation email version
Suj Kang
 
A2 Media Studies - Digipack Powerpoint
A2 Media Studies - Digipack PowerpointA2 Media Studies - Digipack Powerpoint
A2 Media Studies - Digipack Powerpoint
Pheebs023
 
100 formas de animar grupos juegos para usar en talleres reuniones y la comun...
100 formas de animar grupos juegos para usar en talleres reuniones y la comun...100 formas de animar grupos juegos para usar en talleres reuniones y la comun...
100 formas de animar grupos juegos para usar en talleres reuniones y la comun...
steffany Rodriguez
 

Viewers also liked (9)

Sol amortiguadoras
Sol amortiguadorasSol amortiguadoras
Sol amortiguadoras
 
Alexandra Velnidou on Service Design
Alexandra Velnidou on Service DesignAlexandra Velnidou on Service Design
Alexandra Velnidou on Service Design
 
Tugas kkpi
Tugas kkpiTugas kkpi
Tugas kkpi
 
A Personal Testimonial From a Morgan Rothschild Academy Parent
A Personal Testimonial From a Morgan Rothschild Academy ParentA Personal Testimonial From a Morgan Rothschild Academy Parent
A Personal Testimonial From a Morgan Rothschild Academy Parent
 
Cuisine then/now
Cuisine then/nowCuisine then/now
Cuisine then/now
 
All budget-chart-book-2012
All budget-chart-book-2012All budget-chart-book-2012
All budget-chart-book-2012
 
Oil, gas+power customer presentation email version
Oil, gas+power customer presentation   email versionOil, gas+power customer presentation   email version
Oil, gas+power customer presentation email version
 
A2 Media Studies - Digipack Powerpoint
A2 Media Studies - Digipack PowerpointA2 Media Studies - Digipack Powerpoint
A2 Media Studies - Digipack Powerpoint
 
100 formas de animar grupos juegos para usar en talleres reuniones y la comun...
100 formas de animar grupos juegos para usar en talleres reuniones y la comun...100 formas de animar grupos juegos para usar en talleres reuniones y la comun...
100 formas de animar grupos juegos para usar en talleres reuniones y la comun...
 

Similar to Tracking norm

Effective monitoring with StatsD
Effective monitoring with StatsDEffective monitoring with StatsD
Effective monitoring with StatsD
Datadog
 
Wayne State University & DataStax: World's best data modeling tool for Apache...
Wayne State University & DataStax: World's best data modeling tool for Apache...Wayne State University & DataStax: World's best data modeling tool for Apache...
Wayne State University & DataStax: World's best data modeling tool for Apache...
DataStax Academy
 
World’s Best Data Modeling Tool
World’s Best Data Modeling ToolWorld’s Best Data Modeling Tool
World’s Best Data Modeling Tool
Artem Chebotko
 
eXtreme365 - Plugin Development for Analysis
eXtreme365 - Plugin Development for AnalysiseXtreme365 - Plugin Development for Analysis
eXtreme365 - Plugin Development for Analysis
Jonas Rapp
 
Graph processing at scale using spark & graph frames
Graph processing at scale using spark & graph framesGraph processing at scale using spark & graph frames
Graph processing at scale using spark & graph frames
Ron Barabash
 
Patterns of the Lambda Architecture -- 2015 April - Hadoop Summit, Europe
Patterns of the Lambda Architecture -- 2015 April - Hadoop Summit, EuropePatterns of the Lambda Architecture -- 2015 April - Hadoop Summit, Europe
Patterns of the Lambda Architecture -- 2015 April - Hadoop Summit, Europe
Flip Kromer
 
Agile Estimation for Fixed Price Model
Agile Estimation for Fixed Price ModelAgile Estimation for Fixed Price Model
Agile Estimation for Fixed Price Model
jayanth72
 
Data Science & AI Syllabus - DS & AI.pdf
Data Science & AI Syllabus - DS & AI.pdfData Science & AI Syllabus - DS & AI.pdf
Data Science & AI Syllabus - DS & AI.pdf
Aayushdigichrome
 
Python & machine learning Syllabus - DS & AI.pdf
Python & machine learning Syllabus - DS & AI.pdfPython & machine learning Syllabus - DS & AI.pdf
Python & machine learning Syllabus - DS & AI.pdf
Aayushdigichrome
 
Data Science & Artificial intelligence Syllabus - DS & AI.pdf
Data Science & Artificial intelligence Syllabus - DS & AI.pdfData Science & Artificial intelligence Syllabus - DS & AI.pdf
Data Science & Artificial intelligence Syllabus - DS & AI.pdf
DIGICROMESUPPORTTEAM
 
CSE545 sp23 (2) Streaming Algorithms 2-4.pdf
CSE545 sp23 (2) Streaming Algorithms 2-4.pdfCSE545 sp23 (2) Streaming Algorithms 2-4.pdf
CSE545 sp23 (2) Streaming Algorithms 2-4.pdf
AlexanderKyalo3
 
Mike Schaaf -- Synapse ANSYS HFSS Antenna Design
Mike Schaaf -- Synapse ANSYS HFSS Antenna DesignMike Schaaf -- Synapse ANSYS HFSS Antenna Design
Mike Schaaf -- Synapse ANSYS HFSS Antenna Design
Synapse Product Development
 

Similar to Tracking norm (12)

Effective monitoring with StatsD
Effective monitoring with StatsDEffective monitoring with StatsD
Effective monitoring with StatsD
 
Wayne State University & DataStax: World's best data modeling tool for Apache...
Wayne State University & DataStax: World's best data modeling tool for Apache...Wayne State University & DataStax: World's best data modeling tool for Apache...
Wayne State University & DataStax: World's best data modeling tool for Apache...
 
World’s Best Data Modeling Tool
World’s Best Data Modeling ToolWorld’s Best Data Modeling Tool
World’s Best Data Modeling Tool
 
eXtreme365 - Plugin Development for Analysis
eXtreme365 - Plugin Development for AnalysiseXtreme365 - Plugin Development for Analysis
eXtreme365 - Plugin Development for Analysis
 
Graph processing at scale using spark & graph frames
Graph processing at scale using spark & graph framesGraph processing at scale using spark & graph frames
Graph processing at scale using spark & graph frames
 
Patterns of the Lambda Architecture -- 2015 April - Hadoop Summit, Europe
Patterns of the Lambda Architecture -- 2015 April - Hadoop Summit, EuropePatterns of the Lambda Architecture -- 2015 April - Hadoop Summit, Europe
Patterns of the Lambda Architecture -- 2015 April - Hadoop Summit, Europe
 
Agile Estimation for Fixed Price Model
Agile Estimation for Fixed Price ModelAgile Estimation for Fixed Price Model
Agile Estimation for Fixed Price Model
 
Data Science & AI Syllabus - DS & AI.pdf
Data Science & AI Syllabus - DS & AI.pdfData Science & AI Syllabus - DS & AI.pdf
Data Science & AI Syllabus - DS & AI.pdf
 
Python & machine learning Syllabus - DS & AI.pdf
Python & machine learning Syllabus - DS & AI.pdfPython & machine learning Syllabus - DS & AI.pdf
Python & machine learning Syllabus - DS & AI.pdf
 
Data Science & Artificial intelligence Syllabus - DS & AI.pdf
Data Science & Artificial intelligence Syllabus - DS & AI.pdfData Science & Artificial intelligence Syllabus - DS & AI.pdf
Data Science & Artificial intelligence Syllabus - DS & AI.pdf
 
CSE545 sp23 (2) Streaming Algorithms 2-4.pdf
CSE545 sp23 (2) Streaming Algorithms 2-4.pdfCSE545 sp23 (2) Streaming Algorithms 2-4.pdf
CSE545 sp23 (2) Streaming Algorithms 2-4.pdf
 
Mike Schaaf -- Synapse ANSYS HFSS Antenna Design
Mike Schaaf -- Synapse ANSYS HFSS Antenna DesignMike Schaaf -- Synapse ANSYS HFSS Antenna Design
Mike Schaaf -- Synapse ANSYS HFSS Antenna Design
 

Tracking norm

  • 1. Real-time tracking Examples Which norm should we use? Real-time Tracking with Features Petter Strandmark Lund University February 28, 2009 Petter Strandmark Real-time Tracking with Features
  • 2. Real-time tracking Examples Which norm should we use? Outline 1 Real-time tracking SIFT (review) Estimating transformations RANSAC (review) SURF 2 Examples Successful examples Failures Multiple objects 3 Which norm should we use? Line fitting Panorama stitching Petter Strandmark Real-time Tracking with Features
  • 3. Real-time tracking Examples Which norm should we use? SIFT (review) Estimating transformations RANSAC (review) SURF SIFT Petter Strandmark Real-time Tracking with Features
  • 4. Real-time tracking Examples Which norm should we use? SIFT (review) Estimating transformations RANSAC (review) SURF Estimating transformations Petter Strandmark Real-time Tracking with Features
  • 5. Real-time tracking Examples Which norm should we use? SIFT (review) Estimating transformations RANSAC (review) SURF RANSAC RAndom SAmple Consensus Least−squares Least−squares RANSAC Petter Strandmark Real-time Tracking with Features
  • 6. Real-time tracking Examples Which norm should we use? SIFT (review) Estimating transformations RANSAC (review) SURF SURF Gaussian Box filter Gaussian Box filter Approximate Gaussians with box filters. Petter Strandmark Real-time Tracking with Features
  • 7. Real-time tracking Examples Which norm should we use? Successful examples Failures Multiple objects “David” Tracking example Petter Strandmark Real-time Tracking with Features
  • 8. Real-time tracking Examples Which norm should we use? Successful examples Failures Multiple objects Tracking vs. recognition Recognition Tracking Petter Strandmark Real-time Tracking with Features
  • 9. Real-time tracking Examples Which norm should we use? Successful examples Failures Multiple objects “All-Bran” Result Petter Strandmark Real-time Tracking with Features
  • 10. Real-time tracking Examples Which norm should we use? Successful examples Failures Multiple objects “Pedestrian” Result Petter Strandmark Real-time Tracking with Features
  • 11. Real-time tracking Examples Which norm should we use? Successful examples Failures Multiple objects Rotation Is handled well Petter Strandmark Real-time Tracking with Features
  • 12. Real-time tracking Examples Which norm should we use? Successful examples Failures Multiple objects Illumination SIFT descriptors are invariant to illumination Petter Strandmark Real-time Tracking with Features
  • 13. Real-time tracking Examples Which norm should we use? Successful examples Failures Multiple objects Occlusions Only a subset of the SIFT descriptors need to be found Petter Strandmark Real-time Tracking with Features
  • 14. Real-time tracking Examples Which norm should we use? Successful examples Failures Multiple objects Untextured Objects Unable to extract SIFT features Petter Strandmark Real-time Tracking with Features
  • 15. Real-time tracking Examples Which norm should we use? Successful examples Failures Multiple objects Multiple objects Without any modification Petter Strandmark Real-time Tracking with Features
  • 16. Real-time tracking Examples Which norm should we use? Successful examples Failures Multiple objects Multiple objects Additional considerations Petter Strandmark Real-time Tracking with Features
  • 17. Real-time tracking Examples Which norm should we use? Successful examples Failures Multiple objects Multiple objects Only add points within a single frame when updating Petter Strandmark Real-time Tracking with Features
  • 18. Real-time tracking Examples Which norm should we use? Line fitting Panorama stitching Which norm should we use? Petter Strandmark Real-time Tracking with Features
  • 19. Real-time tracking Examples Which norm should we use? Line fitting Panorama stitching Which norm should we use? || · ||∞ Petter Strandmark Real-time Tracking with Features
  • 20. Real-time tracking Examples Which norm should we use? Line fitting Panorama stitching Which norm should we use? || · ||2 Petter Strandmark Real-time Tracking with Features
  • 21. Real-time tracking Examples Which norm should we use? Line fitting Panorama stitching Which norm should we use? || · ||1 Petter Strandmark Real-time Tracking with Features
  • 22. Real-time tracking Examples Which norm should we use? Line fitting Panorama stitching Which norm should we use? Petter Strandmark Real-time Tracking with Features
  • 23. Real-time tracking Examples Which norm should we use? Line fitting Panorama stitching Which norm should we use? || · ||∞ Petter Strandmark Real-time Tracking with Features
  • 24. Real-time tracking Examples Which norm should we use? Line fitting Panorama stitching Which norm should we use? || · ||2 Petter Strandmark Real-time Tracking with Features
  • 25. Real-time tracking Examples Which norm should we use? Line fitting Panorama stitching Which norm should we use? || · ||1 Petter Strandmark Real-time Tracking with Features
  • 26. Real-time tracking Examples Which norm should we use? Line fitting Panorama stitching Resources Course home page SURFmex : http://users.student.lth.se/p04pst/surfmex.html YALMIP optimization package : http://control.ee.ethz.ch/~joloef/wiki/pmwiki.php Petter Strandmark Real-time Tracking with Features