Broadcast News Understanding and Navigation

摘要

The Broadcast News Editor (BNE) and Broadcast News Navigator (BNN) are fully implemented systems that exploit integrated image, speech, and language processing to support intelligent access to broadcast news video. This paper summarizes the integration of a range of AI techniques within these systems to provide intelligent segmentation, extraction, search, summarization, visualization, and personalized multimedia generation. BNE and BNN transform video access from sequential to direct search, providing novel navigation and discovery mechanisms such as topical and entity-specific news clusters. We report an empirical evaluation that demonstrates users can find stories and answer questions nearly three times as fast as searching digital video with no loss in precision and recall.