Home

About

Advanced Search

Browse by Discipline

Scientific Societies

E-print Alerts

Add E-prints

E-print Network
FAQHELPSITE MAPCONTACT US


  Advanced Search  

 
V. Sparse model matching global V.1 Introduction
 

Summary: V. Sparse model matching global
methods
V.1 Introduction
ffl Sparse models matched directly to data, no initialization required.
ffl Data models point processes calculated using binary local features.
ffl Local features photometric and geometric invariant estimated from training
examples in a robust way.
ffl Each landmark in the object has characteristic local topography. See figure V.1
(left). Level curves around left ventricle in an axial MRI image. Characterize
topography with photometric invariant local feature.
ffl Impossible to find feature which only flags correct location. See level curves in
a neighborhood of a sulcus figure V.1 (right).
ffl Find feature with low density of hits away from the correct location on the
object.
ffl One local feature insufficient to detect object: many false positives, minimal
instantiation information use a collection of points associated with a set of
points Z on the reference grid R.
40

V. SPARSE MODEL MATCHING GLOBAL METHODS 41

  

Source: Amit, Yali - Departments of Computer Science & Statistics, University of Chicago

 

Collections: Computer Technologies and Information Sciences