VOISE

Version 1.0.1 - released 23rd March 2011
Developed by Dr. Patrick Guio

VOronoi Image SEgmentation (VOISE) is a dynamic algorithm for partitioning the underlying pixel grid of an image into regions according to a prescribed homogeneity criterion and is presently written in Matlab.

A description of VOISE (including the parameters of each phase as defined below) is available in the following paper Guio, P. and N. Achilleos, The VOISE algorithm: a versatile tool for automatic segmentation of astronomical images, MNRAS 398 1254-1262, 2009 DOI 10.1111/j.1365-2966.2009.15218.x

The paper is available at MNRAS (http://www3.interscience.wiley.com/journal/122540154/abstract) and a free version is also available at arxiv (http://fr.arxiv.org/abs/0906.1905).

It is designed to be web-available on demand as part of the Europlanet RI JRA3 project. Europlanet should be credited for use

Please use the form below to configure and submit your file to VOISE; please note that all fields must be filled in. When the run has completed the results will be emailed to you.

VOISE Parameters

Initialising phase

Number of initial seeds:
Initial value for random number generator:

Dividing phase

Percentile p_D for division: Negative value means no division.
Min square distance between seeds:

Merging phase

Percentile p_M for merging: Negative value means no merging.
Maximum dissimilarity Δμ:
Max ratio non homogeneous/total length H:

Regularisation phase

Maximum numbers of iterations: Negative or null value means no.

Upload File upload / SAMP

This web version of VOISE supports file uploads in the matlab .mat format (the file should contain an image as a 2 dimensional array, labelled as 'Z') and in the .fits format.

File to upload:

SAMP

Image Parameters

Image Origin

Image Origin x:
Image Origin y:

Image Pixel Size

Pixel Size x:
Pixel Size y:

Pixel Units

Pixel Unit x:
Pixel Unit y:

Submit VOISE Job

Email Address: This email address will be used to send you the results.