Summary: IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. 10, NO. 6, JUNE 2001 885
Oriented Texture Completion by AM­FM
Reaction­Diffusion
Scott T. Acton, Senior Member, IEEE, Dipti Prasad Mukherjee, Joebob P. Havlicek, Senior Member, IEEE, and
Alan Conrad Bovik, Fellow, IEEE
Abstract--We provide an automated method to repair broken,
occluded oriented image textures. Our approach is based on par-
tial differential equations (PDEs) and AM­FM image modeling.
Reconstruction of the texture occurs via simultaneous PDE-gen-
erated diffusion and reaction. In the diffusion process, the image
is adaptively smoothed, preserving important boundaries and fea-
tures. The reaction process produces the reconstructed textural
information in the occluded image regions. Gabor filters are de-
signed and used in the reaction process using an AM­FM domi-
nant component analysis. An AM­FM model of the texture image
is constructed, making it possible to localize the reaction filters
spatio­spectrally. In contrast to previous disocclusion techniques
that depend on interpolation, on continuity of the connected com-
ponents within the image level sets, or on texture estimation, the
reaction­diffusion process proposed here yields a seamless transi-

Source: Acton, Scott - Department of Electrical and Computer Engineering, University of Virginia
Havlicek, Joebob - School of Electrical and Computer Engineering, University of Oklahoma

Collections: Computer Technologies and Information Sciences