High-Resolution Image Reconstruction with the ICM Algorithm

Homem, M. R. P.Martins, A. L. D.Mascarenhas, N. D. A.

In this work, we consider the problem of resolutionenhancement from a set of blurred and noisy lowresolutionimages with different sub-pixel displacementsfrom each other. We propose a new procedurefor high-resolution image reconstruction basedon both the Iterated Conditional Modes (ICM) algorithmand a new approach for sub-pixel image registration.The first step is given by aligning all thelow-resolution observations over a high-resolutiongrid and then improve the resolution through theICM algorithm, where a Potts-Strauss model is assumedfor the a priori probability density functionof the actual image. The method was analyzed consideringa number of globally translated and lowresolutionobservations and the results demonstratethe power of the algorithm in reconstructing a highresolutionimage.

Caso o link acima esteja inválido, faça uma busca pelo texto completo na Web: Buscar na Web

Biblioteca Digital Brasileira de Computação - Contato:
     Mantida por: