FIND: facilitating the identification of intervals and moments for incorporation of additional content in continuous media

• Kamila R. H. Rodrigues UFSCar
• Maria da Graça Pimentel USP
• Suetônio S. Pereira UFSCar
• Cesar A. C. Teixeira UFSCar

Resumo

As a result of many real world demands, the literature abounds with results in the form of annotation-based systems, as it is the case of systems which allow the annotation of audio or video objects as well as digital ink, paper and web pages, to name a few. Among the many types of annotations which may be of interest, annotations in the form of additional content associated with audio or video objects are likely to result in enriched media useful for several applications, including the promotion of accessibility. However, besides the highly demanding intellectual work of defining what additional content should be created for a particular segment of media, the author of that content faces the task of identifying the exact media intervals in which the additional content should be inserted. That task is particularly arduous when the media has a long duration. In this paper we propose FIND, a reference model guiding the construction of systems which aggregate software components for: 1) the specification and the recognition of patterns in audio or video; 2) the synchronized composition of the annotated media; 3) and the formatting of the annotated media into multimedia documents for presentation. We also illustrate how the model was instantiated in a configuration which allows the association of additional content to movies resulting in a multimedia document with better accessibility.