Texture Analysis of Painted Strokes (bibtex)
by Martin Lettner, Paul Kammerer, Robert Sablatnig
Abstract:
In this practical work texture analysis for painted strokes is rep orted. The work presents a study of stroke classification in which two classes of strokes are identified: fluid and dry strokes. The discrimination is done with a feature vector which is extracted from the stroke texture by the help of texture analysis metho ds. To find an adequate texture analysis metho d for this application, three different texture analysis metho ds are executed on test images from painted strokes. The metho ds applied are based on statistical features of first and second order and on the discrete wavelet transformation, whereas the statistical features of second order are extracted from the co-o ccurrence matrix. The results are compared and it turns out that the wavelet based texture analysis metho d yields the b est discrimination rate for this application
Reference:
Texture Analysis of Painted Strokes (Martin Lettner, Paul Kammerer, Robert Sablatnig), Technical report, PRIP, TU Wien, 2004.
Bibtex Entry:
@TechReport{PTR-Lettner04a,
  author =	 "Martin Lettner and Paul Kammerer and Robert
                  Sablatnig",
  title =	 "Texture Analysis of Painted Strokes",
  institution =	 "PRIP, TU Wien",
  number =	 "PRIP-TR-089",
  year =	 "2004",
  url =		 "ftp://ftp.prip.tuwien.ac.at/pub/publications/trs/tr89.pdf",
  abstract =	 " In this practical work texture analysis for painted
                  strokes is rep orted. The work presents a study of
                  stroke classification in which two classes of
                  strokes are identified: fluid and dry strokes. The
                  discrimination is done with a feature vector which
                  is extracted from the stroke texture by the help of
                  texture analysis metho ds. To find an adequate
                  texture analysis metho d for this application, three
                  different texture analysis metho ds are executed on
                  test images from painted strokes. The metho ds
                  applied are based on statistical features of first
                  and second order and on the discrete wavelet
                  transformation, whereas the statistical features of
                  second order are extracted from the co-o ccurrence
                  matrix. The results are compared and it turns out
                  that the wavelet based texture analysis metho d
                  yields the b est discrimination rate for this
                  application ",
}
Powered by bibtexbrowser