| Abstract: |
| In this talk, we deal with the application of Besov spaces in image processing. Although Besov images has been already studied in the literature, we treat the problem with different methods (fractal-wavelet modeling, calculus of variations, etc.) in order to give a recent overview with new results. First, the norm equivalence between $L^p$ and $\ell^p$ allows the wavelet characterization of Besov images. We point out that the sparsity of the wavelet representation directly leads us to investigate the compression image in terms of approximation error. Secondly, the calculus of variations allows us to improve the process of denoising together with some results in terms of compression image. Moreover, we generalize these results in other function spaces. Finally, we give some examples of the results above through the introduction of wavelet families. |
|