Semantic Approach for the Spatial Adaptation of Multimedia Documents

Abstract

With the proliferation of heterogeneous devices (desktop computers, personal digital assistants, phones), multimedia documents must be played under various constraints (small screens, low bandwidth, etc). Taking these constraints into account with current document models is impossible. Hence, generic source documents must be adapted or transformed into documents compatibles with the target contexts. The adaptation consists of modifying this specification of the document in a minimal way to lead it to satisfy the target profile. The profile defines constraints that must be satisfied by the document to be played. At this level, the transgressive adaptation becomes necessary when no specification model exists to satisfy this profile. We focus on the spatial dimension of a multimedia document and we provide an approach for the spatial adaptation that permits to best preserve the initial document semantic. In the proposed transgressive adaptation, the relations that do not comply with the target profile are not most often replaced by the closest ones due to the fact that their immediate neighboring relations have the same similarity degree whereas there may exist differences between them. In this paper, we extend this approach to, firstly, best preserve the proximity between the adapted and the original documents by weighting the arcs of the conceptual neighborhood graphs and secondly, to deal with complex relations models by integrating the concept of the relations relaxation graphs that permit to handle the distances defined within the relations.