2013 | OriginalPaper | Buchkapitel
A Graph-Cut-Based Smooth Quantization Approach for Image Compression
verfasst von : Maria Trocan, Beatrice Pesquet
Erschienen in: Computational Collective Intelligence. Technologies and Applications
Verlag: Springer Berlin Heidelberg
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
Quantization represents an important aspect in image acquisition and coding. However, the classical quantization algorithms lack of spatial smoothness, especially when dealing with low bitrate constraints. In this paper, we propose a graph-cut-based smooth quantization approach for image compression that can alleviate the artefacts driven by classical quantization algorithms. The best representation for an image using a finite number of levels is obtained by convex optimization, realized by graph-cut techniques and which considers the spatial correlation in the minimization process in addition to the classical distortion approach. We show that even when using a small number of reconstruction levels, our approach can yield better quality results, in terms of PSNR, than JPEG2000.