The problem of image restoration is considered when the point spread function is Space Variant Non Separable. The algorithm determines a continuous approximation of the original object, following the continuous object-discrete image approach. The image spatial domain is decomposed into subdomains and the local approximants are computed on a distributed memory environment. The continuity of the solution across the image subdomains is obtained by adding a suitable overlapping area to the sides of the subdomains. Numerical experiments have been carried out on a Hypercube Intel iPSC/860 and the most interesting results are reported.
Baronio A., Zama F. (1996). A domain decomposition technique for spline image restoration on distributed memory systems. PARALLEL COMPUTING, 22(1), 101-110 [10.1016/0167-8191(95)00067-4].
A domain decomposition technique for spline image restoration on distributed memory systems
Zama F.
1996
Abstract
The problem of image restoration is considered when the point spread function is Space Variant Non Separable. The algorithm determines a continuous approximation of the original object, following the continuous object-discrete image approach. The image spatial domain is decomposed into subdomains and the local approximants are computed on a distributed memory environment. The continuity of the solution across the image subdomains is obtained by adding a suitable overlapping area to the sides of the subdomains. Numerical experiments have been carried out on a Hypercube Intel iPSC/860 and the most interesting results are reported.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.