Large-scale infographic image downsizing

´Content repurposing´ is a way to convert existing multimedia content to suit mobile devices (handphones, PDAs, handheld PCs etc.), which have a lower display capability. As long as images are concerned, the main task is to downscale images and images of different types must be tackled differently in order to preserve as much visual content as possible. While in general most images stand fairly well downscaling, infographics (charts, maps, etc.) and other images containing thin structures are delicate to handle. Conventional methods tend to average out thin structures which usually convey critical information. Specifically designed techniques present various kinds of limitations. In this paper, we present a novel approach for infographic image downsizing. It relies on the concept of thin structure´, the determination of which is related to the required downsizing scale. The algorithm is general, robust, without any limitation on the images. It is efficient because all operations are local. It works in an iterative way to achieve arbitrary large-scale downsizing. Result examples on real-world images are given to show the effectiveness of the approach.