Title :
A Volterra type model for image processing
Author :
Cottet, Georges Henri ; Ayyadi, M.E.
Author_Institution :
LMC-IMAG, Univ. de Grenoble, France
fDate :
3/1/1998 12:00:00 AM
Abstract :
We present a class of time-delay anisotropic diffusion models for image restoration. These models lead to asymptotic states that are selected on the basis of a contrast parameter and bear some analogy with neural networks with Hebbian dynamical learning rules. Numerical examples show that these models are efficient in removing even high levels of noise, while allowing an accurate tracking of the edges
Keywords :
Hebbian learning; Volterra equations; delays; edge detection; image restoration; image segmentation; neural nets; noise; parameter estimation; partial differential equations; Hebbian dynamical learning rules; Volterra type model; asymptotic states; contrast parameter; edge tracking; image processing; image restoration; image segmentation; neural networks; noise removal; nonlinear diffusion model; partial differential equations; time-delay anisotropic diffusion models; Acoustic noise; Adaptive systems; Anisotropic magnetoresistance; Equations; Filters; Image processing; Image restoration; Mathematical model; Neural networks; Noise level;
Journal_Title :
Image Processing, IEEE Transactions on
