Automatic detection of flamingos using marked point processes
Click here to begin the demo.
For more information on this research topic see the INRIA research report (here).
To get the code of FLAMINGO software (CeCILL C) send a mail to J. Zerubia (Josiane.Zerubia@sophia.inria.fr).
Abstract
Here we present a new technique to automatically detect and count breeding Greater
flamingos (Phoenicopterus roseus) on aerial photographs of their colonies. We consider a stochastic
approach based on object processes also called marked point processes. Here, the objects represent
flamingos which are defined as ellipses. The density associated with the marked point process of
ellipses is defined w.r.t the Poisson measure. Thus, the issue is reduced to an energy minimization,
where the energy is composed of a regularization term (prior density), which introduces some
constraints on the objects and their interactions, and a data term, which links the objects to the
features to be extracted in the image. Then, we sample the process to extract the configuration of
objects minimizing the energy by a new fast birth-and-death dynamics, leading to the total number
of birds. This approach gives counts with good precision compared to manual counts. Additionally,
this approach does not need image pre-processing or supervision of the extraction thus considerably
reducing the overall processing time required to get the estimate.
Acknowledgement
We would like to thank the foundation Tour Du Valat, Arles, France, for providing the data. This work has been partially funded by
the INRIA Sophia Antipolis with a COLOR "Flamants".
Stig Descamps
2008-03-20