Perception-driven resizing for dynamic image sequences

With the development of multimedia display devices, dynamic image sequence resizing, which can adapt image sequences to be displayed on devices with different resolutions, is becoming more important. However, existing approaches do not resize results from the viewpoint of the user. In this paper, we present a new resizing framework, which uses the feature descriptor technique and the image interpolation technique, that aims to improve the resizing quality of the important content perceived by user. To accomplish this, we use a coarse-to-fine detection approach to determine the important content of image sequences, and construct a partition interpolation model to improve the definitions of important content. By adopting a region energy protection approach we can obtain high quality image displays. Compared to representative algorithms in image resizing, our method can achieve satisfactory performance not only in terms of image visualization but also in terms of quantitative measures.