-The multilayer perceptron has a large wide of classification and regression applications in many fields: pattern recognition, voice and classification problems. But the architecture choice has a great impact on the convergence of these networks. In the present paper we introduce a new approach to optimize the network architecture, for solving the obtained model we use the genetic algorithm and we train the network with a back-propagation algorithm. The numerical results assess the effectiveness of the theoretical results shown in this paper, and the advantages of the new modeling compared to the previous model in the literature.

No abstract

No abstract

No abstract

I. INTRODUCTION U Ncompressed multimedia data require considerable storage capacity and transmission bandwidth. Thus, image compression is a very important factor for better utilization of network bandwidth and computer storage. The compression process is usually lossy and is based on redundancy and irrelevancy reduction, which are inherent in the image domain. In the medical field, using radiographic images, ultrasound, MRI (Magnetic Resonance Imaging), ... poses a great problem of storage and archiving. To overcome these problems, compression of these images is an operation necessary and imperative. The main purpose of image compression is to reduce the amount of bits needed to describe them while keeping an acceptable visual appearance of the reconstructed images. The compression of images has been performed by several techniques among the best known: the JPEG is a lossy method standardized by ISO in August 1990, these methods perform compression by performing a scalar quantization (SQ) on the values obtained after processing. The disadvantage of the scalar quantization is that it does not exploit the spatial correlation between different pixels of the image. Another more interesting way to achieve compression coding is not the values individually one after the other, but to encode a set of values simultaneously. This procedure is called vector quantization (VQ) [12], the VQ has been successfully used for encoding the voice signal and for compressing still images. Approaches using artificial neural networks for intelligent