Extensions of invariant signatures for object recognition

Daniel J. Hoff, Peter J. Olver

Research output: Contribution to journalArticlepeer-review

27 Scopus citations

Abstract

A refinement of the method of differential invariant signatures for object recognition is presented. The value of the method lies in its compromise between local and global identifying properties, thereby allowing us to distinguish non-congruent curves whose Euclidean signatures have identical trace.

Original languageEnglish (US)
Pages (from-to)176-185
Number of pages10
JournalJournal of Mathematical Imaging and Vision
Volume45
Issue number2
DOIs
StatePublished - Feb 2013

Bibliographical note

Funding Information:
Supported in part by NSF Grant DMS 08–07317.

Keywords

  • Bivertex arc
  • Curvature
  • Differential invariant
  • Euclidean signature
  • Object recognition
  • Plane curve
  • Rigid equivalence
  • Vertex

Fingerprint

Dive into the research topics of 'Extensions of invariant signatures for object recognition'. Together they form a unique fingerprint.

Cite this