Integration of a human face annotation technology in an audio-visual search engine platform

Abstract

Search-Based Web Applications are configurable and customizable platforms for multimedia search over the Web. This paper describes the results of the integration of an annotation component for automatic face recognition within a configurable multimedia Web search platform. We integrate a third-party image analysis technology for human face recognition, by extending it in order to cope with the requirements of video annotation solutions on the Web and wrapping it according to the platform's integration specification. As a result, the annotator outcomes are indexed within the platform and can be used for multi-modal and content-based search. We evaluate our experience in terms of development effort required, indexing performances, and perceived quality of the annotation results.