Total Variation Based Oversampling of Noisy Images

Abstract. We propose a variational model which permits to simultaneously deblur and oversample an image. Indeed, after some recalls on an existing variational model for image oversampling, we show how to modify it in order to properly achieve our two goals. We discuss the modification both under a theoretical point of view (the analysis of the preservation of some structural elements) and the practical point of view of experimental results. We also describe the algorithm used to compute a solution to this model.

• Type: Conference
• Year: 2001
• Where: SCALESPACE
• Authors: François Malgouyres
• Keywords: Computer Vision, SCALESPACE 2001, Structural Elements, Theoretical Point, Variational Model