pnonlocalerosion

Synopsis

Description

Performs a non local erosion on image im_in. Let $f$ be im_in, this algorithm performs iteratively number_of_iterations times:
$ f(u)^{t+1} = f^t(u) + \min_{v ~ u}(w(u, v) \min(f^t(v) - f^t(u), 0))$
with $u$ a pixel location, $v$ a neighbor of $u$, $w(u, v)$ the weight between $u$ and $v$ (which can be a similarity measure, a distance, etc.).

Here we use the exponential similarity measure: $w(u,v)=\exp{ -distance(f(u), f(v))^2 / sigma^2 }$. If you use $w(u, v) = 1$, you get the typical erosion operator.

Parameters

• sigma: sigma used for the similarity measure.
• number_of_iterations: number of iterations.
• connectivity: neighborhood, can be 4 or 8.

Inputs

• im_in: a 2D image.

Outputs

• im_out: an image of the same type as the input image.

Result

Returns SUCCESS or FAILURE.

Examples

• Apply a non local erosion to tangram.pan, with 8-connexity neighborhood, 10 iterations, and sigma = 15.

  pnonlocalerosion 15 10 8 tangram.pan out.pan
  

See also

C++ prototype