An affine shape constraint for geometric active contours - Emerging Trends in Image Processing, Computer Vision, and Pattern Recognition - page 433

clutered objects based on the presented aine shape alignment method for the purpose of a
good segmentation result ( Figure 13 ) .
FIGURE 13 Robust object detection in mosaic image.
5 Conclusions
We presented in this paper an alternative approach to incorporate prior knowledge into a level
set-based active contours in order to have robust object detection in case of large shape dis-
tortion that can be analyzed by the class of affine transformations. We presented also a geo-
metric solution to choose the reference shape in case of many available templates, given that
the statistical approach needs a training set and PCA. Then the application of a given classifier
like Bayesian classifier to determine the appropriate reference shape like the work of Fang and
Chan [ 24 ] . Given that the proposed approach invokes only pixels of the regions of variability
between the shape of reference and the object of interest in the process of curve evolution and
based on the fast Fourier transform for affine motion estimation and invariants computation,
the method is faster compared to Refs. [ 15 , 25 ] where at each iteration shape descriptors are
calculated for a given order that has to be set empirically. The obtained results are promising
in the case of real and simulated data and the method can be used for the restoration of mosaic
images in the archeological field. As future perspectives, we are working on integrating our
model in the context of 3D object reconstruction from silhouetes sequence in order to reine
the obtained 3D model.
References
[1] Kass M, Witkin A, Terzopoulos D. Snakes: active contour models. Int J Comput Vis.
1988;1:321-331.
[2] Osher S, Sethian JA. Fronts propagating with curvature-dependent speed: algorithms
based on Hamilton-Jacobi formulation. J Comput Phys. 1988;79:12-49.
[3] Caselles V, Kimmel R, Sapiro G. Geodesic active contours. Int J Comput Vis.
1997;22:61-79.
[4] Malladi R, Sethian J, Vemuri B. Shape modeling with front propagation: a level set
approach. PAMI. 1995;17:158-175.
[5] Chan T, Vese L. Active contours without edges. IEEE Trans Imag Process.
2001;10:266-277.
[6] Charmi MA, Mezghich MA, M'Hiri S, Derrode S, Ghorbel F. Geometric shape prior
to region-based active contours using Fourier-based shape alignment. In: IEEE international
conference on imaging systems and techniques; 2010:478-481.