Iconic representation for progressive transmission of medical images

Abstract

Progressive image transmission (PIT) has been shown to be effective in providing timely delivery of medical images over limited bandwidth communication channels. A novel method for identifying regions of interest and producing iconic caricatures is presented. The generation of these caricatures is designed as a first step in a context-based form of progressive image transmission. Regions of interest are identified but images are not segmented. The identification of regions of interests uses shape information provided by multi-scale medial axis (MMA). This approach avoids the need for prior knowledge to label regions of an image. An approximate labelling is used to create iconic images. The results presented show improved ability to detect low contrast regions and delineation appropriate to the context to provide a good caricature. These results confirm that robust and reliable iconic representation can be generated without prior knowledge of image content. The method described also has potential for extension to colour and 3D image data sets.