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 ScienceWISE logo

Computer Science > Computer Vision and Pattern Recognition

Title: Comparison between EM and FCM algorithms in skin tone extraction

Abstract: This study aims to investigate implementing EM and FCM algorithms for skin color extraction. The capabilities of three well-known color spaces, namely, RGB, HSV, and YCbCr for skin-tone extraction are assessed by using statistical modeling of skin tones using EM and FCM algorithms. The results show that utilizing a Gaussian mixture model for parametric modeling of skin tones using EM algorithm works well in HSV color space when all three components of the color vector are used. In spite of discarding the luminance components in YCbCr and HSV color spaces, EM algorithm provides the best results. The results of the detailed comparisons are explained in the conclusion.
Comments: 2016 1st International Conference on New Research Achievements in Electrical and Computer Engineering (ICNRAECE)
Subjects: Computer Vision and Pattern Recognition (cs.CV); Image and Video Processing (eess.IV)
Cite as: arXiv:2211.09979 [cs.CV]
  (or arXiv:2211.09979v1 [cs.CV] for this version)

Submission history

From: Elham Ravanbakhsh [view email]
[v1] Fri, 18 Nov 2022 01:53:51 GMT (2308kb)

Link back to: arXiv, form interface, contact.