TRECVID 2017, 21th International Workshop on Video Retrieval Evaluation, November 13-15, 2017, Gaithersburg, USA
This paper describes the system we designed to address the Hyperlinking task at TRECVID 2017 and the achieved results. Our contribution explores the potential of a solution based on the combination of textual and visual features in order to consider the di erent facets of the input videos. In particular, our approaches combined automatically generated transcripts (LIMSI), visual concepts, Meta-data, the text extracted by means of a Name Entity Recognition technique and a concept mapping tool. The four submitted runs aimed at analyzing the impact of the considered features on the quality of the retrieved hyperlinks.
Type:
Conférence
City:
Gaithersburg
Date:
2017-11-13
Department:
Data Science
Eurecom Ref:
5381
Copyright:
© NIST. Personal use of this material is permitted. The definitive version of this paper was published in TRECVID 2017, 21th International Workshop on Video Retrieval Evaluation, November 13-15, 2017, Gaithersburg, USA and is available at :
See also:
