Weighted optimum bit allocation for multiresolution satellite image coding
Marc Antonini, Michel Barlaud, Bernard Rougé, Catherine Lambert-Nebout · 2002
This paper develops a new optimal bit allocation algorithm for efficient transmission of vector sources over a digital noiseless channel. This method takes into account the characteristics of the compression scheme. The algorithm is based on the minimization of a non-convex criterion. In fact, the introduction of constraints, such as positivity constraint, could provide a non-convex criterion. Then, the optimization method may be caught in local minima. To solve this problem, we propose to use the augmented Lagrangian method. We introduce a penalty function which ensures the existence of a saddle-point and the convergence of the algorithm. Then, the minimization problem can be solved using the classical conjugate gradient method.>