View source for Publications/carlinet.15.itip

{{Publication
| published = true
| date = 2015-10-26
| authors = Edwin Carlinet, Thierry Géraud
| title = MToS: A Tree of Shapes for Multivariate Images
| journal = IEEE Transactions on Image Processing
| volume = 24
| number = 12
| pages = 5330 to 5342
| lrdeprojects = Olena
| abstract = The Tree of Shapes (ToS) is a morphological tree that provides an high-level hierarchical representation of the image suitable for many image processing tasks. When dealing with color images, one cannot use the ToS because its definition is ill-formed on multivariate data. Common workarounds such as marginal processing, or imposing a total order on data are not satisfactory and yield many problems (color artifacts, loss of invariances...) In this paper, we highlight the need for a self-dual and contrast invariant representation of the image and provide a method that builds a single ToS by merging the shapes computed marginally and preserving the most important properties of the ToS. This method does not try to impose an arbitrary total ordering on values but uses only the inclusion relationship between shapes and the merging strategy works in a shape space. Eventually, we show the relevance of our method and our structure through several applications involving color and multispectral image analysis.
| lrdeinc = Publications/carlinet.15.itip.inc
| lrdekeywords = Image
| lrdepaper = https://www.lrde.epita.fr/dload/papers/carlinet.15.itip.pdf
| lrdenewsdate = 2015-10-26
| None = 10.1109/TIP.2015.2480599
| type = article
| id = carlinet.15.itip
| identifier = doi:10.1109/TIP.2015.2480599
| bibtex = 
@Article<nowiki>{</nowiki>	  carlinet.15.itip,
  author	= <nowiki>{</nowiki>Edwin Carlinet and Thierry G\'eraud<nowiki>}</nowiki>,
  title		= <nowiki>{</nowiki><nowiki>{</nowiki>MToS<nowiki>}</nowiki>: A Tree of Shapes for Multivariate Images<nowiki>}</nowiki>,
  journal	= <nowiki>{</nowiki>IEEE Transactions on Image Processing<nowiki>}</nowiki>,
  year		= 2015,
  volume	= <nowiki>{</nowiki>24<nowiki>}</nowiki>,
  number	= <nowiki>{</nowiki>12<nowiki>}</nowiki>,
  pages		= <nowiki>{</nowiki>5330--5342<nowiki>}</nowiki>,
  month		= dec,
  abstract	= <nowiki>{</nowiki>The Tree of Shapes (ToS) is a morphological tree that
		  provides an high-level hierarchical representation of the
		  image suitable for many image processing tasks. When
		  dealing with color images, one cannot use the ToS because
		  its definition is ill-formed on multivariate data. Common
		  workarounds such as marginal processing, or imposing a
		  total order on data are not satisfactory and yield many
		  problems (color artifacts, loss of invariances...) In this
		  paper, we highlight the need for a self-dual and contrast
		  invariant representation of the image and provide a method
		  that builds a single ToS by merging the shapes computed
		  marginally and preserving the most important properties of
		  the ToS. This method does not try to impose an arbitrary
		  total ordering on values but uses only the inclusion
		  relationship between shapes and the merging strategy works
		  in a shape space. Eventually, we show the relevance of our
		  method and our structure through several applications
		  involving color and multispectral image analysis.<nowiki>}</nowiki>,
  url		= <nowiki>{</nowiki>10.1109/TIP.2015.2480599<nowiki>}</nowiki>,
  doi		= <nowiki>{</nowiki>10.1109/TIP.2015.2480599<nowiki>}</nowiki>
<nowiki>}</nowiki>

}}