喜綿 洋人

The hyperparameter in image restoration by the Bayes formula is an important quantity. This communication shows a physical method for the estimation of the hyperparameter without approximation. For artificially generated images by prior probability, the hyperparameter is computed accurately. For practical images, accuracy of the estimated hyperparameter depends on the magnetization and energy of the images. We discuss the validity of prior probability for an original image.

We investigate a stability of a lamellar domain in phase-separating binary fluids under an external flow. Using the Navier-Stokes and the Cahn-Hilliard equations, we take into account effects of diffusion and surface tension at an interface. Stability eigenvalues are evaluated for various values of the Peclet number, the spacing between the interfaces, and the Reynolds number. It is found that the lamellar domain becomes unstable at a finite wavenumber before the flow when the Reynolds number increases. The instability of the interface occurs on conditions that the interface is situated near a wall or the Peclet number is large. The instability stems from the interaction between disturbances of the flow and the diffusive interface. (C) 2003 American Institute of Physics.