Event Alignment for Cross-Media Feature Extraction in the Football Domain

This paper describes an experiment in creating cross-media descriptors from football-related text and videos. We used video analysis results and combined them with several textual resources – both semi-structured (tabular match reports) and unstructured (textual minute-by-minute match reports). Our aim was to discover the relations among six video data detectors and their behavior during a time window that corresponds to an event described in the textual data. The experiment shows how football events extracted from text can be mapped to corresponding scenes in video and how this may help in extracting event-specific video detectors.