Focus Based Resource-Aware Semantic Transcoding for Universal Access

Abstract

Users around the world are demanding universal access to network information on all sorts of networks, platforms, device models, and access styles. Traditional transcoding techniques largely center around prior media transcoding targeting predetermined set of network and device configurations without considering user preferences or resource usage. We propose a framework for universal access with focus-based information extraction and resource-aware semantic transcoding. The user focus tree is constructed based on domain ontology reasoning, access behavior monitoring, and focus weight computation. The tree is used for the initial selection of candidate objects for subsequent transcoding. To optimize client device resource usage, the ideas of resource consumption index and information preservation ratio are proposed for the generation of resource-aware semantic transcoding plan which can effectively put the client device resource into best use of producing the right content for the user. Preliminary implementation and evaluation results demonstrate that our techniques can correctly capture user focus and generate quality transcoding plan to achieve high information preservation ratio and high resource usage at the same time.