From metric homotopies to nD From metric homotopies to nD persistence persistence Massimo Ferri Mathematics Department Univ. of Bologna http://www.dm.unibo.it/~ferri/e.htm [email protected]
Mar 28, 2015
From metric homotopies to nD From metric homotopies to nD persistencepersistence
Massimo Ferri
Mathematics Department
Univ. of Bologna
http://www.dm.unibo.it/~ferri/e.htm
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 2/57
From metric homotopies to nD persistenceFrom metric homotopies to nD persistence
• The University of Bologna• Metric homotopies• Size Functions and Natural Pseudodistance• Applications• nD persistence• Work in progress and conclusions
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 3/57
The University of BolognaThe University of Bologna
• Oldest in Europe? (Competitor: La Sorbonne)
• Active at the end of the 11th century as a Law school
• 1158: formal recognition by Emperor Frederick I Barbarossa
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 4/57
The University of BolognaThe University of Bologna
Some “visiting professors”:• Dante Alighieri• Thomas Becket • Erasmus of Rotterdam
A good student:• Nicolaus Copernicus
Thomas Becket
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 5/57
The University of BolognaThe University of BolognaMathematics in Bologna:• Luca Pacioli (14th c.): arithmetic and geometry• Rafael Bombelli (16th c.): invention of complex #’s• Scipione Dal Ferro, Gerolamo Cardano, Ludovico
Ferrari (16th c.): 3rd and 4th degree formulas• Bonaventura Cavalieri, Pietro Mengoli (17th c.):
early integral calculus• Maria Gaetana Agnesi (18th c.): analytical geometry• Luigi Cremona, Eugenio Beltrami, Beniamino Segre
(19th-20th c.): algebraic geometry• Cesare Arzela`, Leonida Tonelli (20th c.): analysis
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 6/57
From metric homotopies to nD persistenceFrom metric homotopies to nD persistence
• The University of Bologna• Metric homotopies• Size Functions and Natural Pseudodistance• Applications• nD persistence• Work in progress and conclusions
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 7/57
Metric homotopiesMetric homotopies
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 8/57
Metric homotopiesMetric homotopies
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 9/57
Metric homotopiesMetric homotopies
Two minimal paths…
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 10/57
Metric homotopiesMetric homotopies
…showing the lack of associativity
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 11/57
Metric homotopiesMetric homotopies
• No way of obtaining a group• Just a fake one…
… which may be fairly complicated.
A minimal path on a cube
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 12/57
Metric homotopiesMetric homotopies
My student Patrizio Frosini decided for a totally different approach:
Size Functions
Frosini, P., Measuring shapes by size functions, Proc. of SPIE, Intelligent Robots and Computer Vision X: Algorithms and Techniques, Boston, MA 1607 (1991).
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 13/57
From metric homotopies to nD persistenceFrom metric homotopies to nD persistence
• The University of Bologna• Metric homotopies• Size Functions and Natural Pseudodistance• Applications• nD persistence• Work in progress and conclusions
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 14/57
Size Functions and Natural PseudodistanceSize Functions and Natural Pseudodistance
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 15/57
Size Functions and Natural PseudodistanceSize Functions and Natural Pseudodistance
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 16/57
Size Functions and Natural PseudodistanceSize Functions and Natural Pseudodistance
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 17/57
Size Functions and Natural PseudodistanceSize Functions and Natural Pseudodistance
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 18/57
Size Functions and Natural PseudodistanceSize Functions and Natural Pseudodistance
Approximation translates into “blind strips”.
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 19/57
Size Functions and Natural PseudodistanceSize Functions and Natural Pseudodistance
All information carried by a size function can be condensed in the formal series of its cornerpoints
The matching distance
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 20/57
Size Functions and Natural PseudodistanceSize Functions and Natural Pseudodistance
The matching distance between formal series of cornerpoints is stable under perturbation of the measuring function!
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 21/57
Size Functions and Natural PseudodistanceSize Functions and Natural Pseudodistance
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 22/57
Size Functions and Natural PseudodistanceSize Functions and Natural Pseudodistance
It turns out that:
i.e. the matching distance between size functions yields a lower bound (and an optimal one!) to the natural pseudodistance.
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 23/57
From metric homotopies to nD persistenceFrom metric homotopies to nD persistence
• The University of Bologna• Metric homotopies• Size Functions and Natural Pseudodistance• Applications• nD persistence• Work in progress and conclusions
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 24/57
ApplicationsApplications
• Classification problems:Bologna
• Leukocytes• Monograms• Sketches• Melanocytic lesions
Genova• Tree leaves• Numerals• Alphabet of the deaf• Cars
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 25/57
ApplicationsApplications
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 26/57
ApplicationsApplications
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 27/57
ApplicationsApplications
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 28/57
ApplicationsApplications
Similitudes
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 29/57
ApplicationsApplications
Affine transformations
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 30/57
ApplicationsApplications
Homographies
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 31/57
ApplicationsApplications
melanoma
naevus
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 32/57
ApplicationsApplications
An image and one of its splittings.
The curve of the image(meas. fct.: luminance).
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 33/57
ApplicationsApplications
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 34/57
ApplicationsApplications
• Image retrievalSea faunaSilhouettesTrade marks
• “Keypics”
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 35/57
ApplicationsApplications
Query 1 2 3
The challenge of a public database.
CSS
CSS
our system
our system
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 36/57
ApplicationsApplicationsQuery 1 2 3
The challenge of a public database.
CSS
CSS
our system
our system
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 37/57
ApplicationsApplications
Trade marks: two queries and the first eight retrieved images
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 38/57
ApplicationsApplications
• We suggest that images on the Internet should be equipped with simplified sketches representing the essentials of the images themselves: keypics.
• Keypics should be provided by the image owner or manager.
• This graphical indexing might be extended to whole Web pages.
• Encoding of keypics should be standard (e.g. in SVG).
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 39/57
ApplicationsApplications
A Data Manager might wish to index the image of a saxophone by its geometrical outline, but also (or only) with a musical note.
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 40/57
ApplicationsApplications
Some different keypic drawing conceptions.
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 41/57
ApplicationsApplications
A retrieval experiment
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 42/57
ApplicationsApplications
A retrieval experiment
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 43/57
From metric homotopies to nD persistenceFrom metric homotopies to nD persistence
• The University of Bologna• Metric homotopies• Size Functions and Natural Pseudodistance• Applications• nD persistence• Work in progress and conclusions
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 44/57
nD persistencenD persistence
• k-dimensional measuring functions• Higher degree homology
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 45/57
nD persistencenD persistence
Frosini, P., Mulazzani, M., Size homotopy groups for computation of natural size distances, Bull. of the Belgian Math. Soc. - Simon Stevin, 6 (1999), 455-464.
k-dimensional measuring functions
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 46/57
nD persistencenD persistence
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 47/57
nD persistencenD persistence
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 48/57
nD persistencenD persistence
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 49/57
nD persistencenD persistence
Higher homology modules (persistent homology / size functor)
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 50/57
nD persistencenD persistence
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 51/57
nD persistencenD persistence
Persistent homology for k-dimensional measuring functions
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 52/57
nD persistencenD persistence
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 53/57
nD persistencenD persistence
The reduction theorem from k-dimensional to 1-dimensional measuring functions - through maxima along admissible pairs - works unaltered also for higher homology modules!
Cagliari, F., Di Fabio, B., Ferri, M., One-dimensional reduction of multidimensional persistent homology, Proc. Amer. Math. Soc. 138 (2010), 3003-3017.
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 54/57
From metric homotopies to nD persistenceFrom metric homotopies to nD persistence
• The University of Bologna• Metric homotopies• Size Functions and Natural Pseudodistance• Applications• nD persistence• Work in progress and conclusions
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 55/57
Work in progress and conclusionsWork in progress and conclusions
Present research:
• Search for optimal admissible pairs• Better bounds for the natural pseudodistance• Algorithms• Applications to colour images and 3D meshes• New applications to melanoma diagnosis
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 56/57
Work in progress and conclusionsWork in progress and conclusions
The theory of Size Functions is developing along two directions: k-dimensionality and higher homology modules.
Together with “concrete” applications in Pattern Recognition, we would like to find contact points with other theoretical domains.
Ljubljana, 12/4/2012 M. Ferri – From metric homotopies to nD persistence 57/57
THANK YOU FOR YOUR ATTENTION !