We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.CV

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo

Computer Science > Computer Vision and Pattern Recognition

Title: Determination of the most representative descriptor among a set of feature vectors for the same object

Abstract: On an example of solution of the face recognition problem the approach for estimation of the most representative descriptor among a set of feature vectors for the same face is considered in present study. The estimation is based on robust calculation of the mode-median mixture vector for the set as the descriptor by means of Welsch/Leclerc loss function application in case of very sparse filling of the feature space with feature vectors
Comments: 8 pages, 1 figure
Subjects: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
Cite as: arXiv:2007.03021 [cs.CV]
  (or arXiv:2007.03021v1 [cs.CV] for this version)

Submission history

From: Dmitry Pozdnyakov [view email]
[v1] Mon, 6 Jul 2020 19:09:06 GMT (607kb)

Link back to: arXiv, form interface, contact.