Implementación de una aplicación web social para el etiquetado de fragmentos de videos de youtube aplicando el concepto de ‘juegos con un propósito’

Abstract

This research presents the design and implementation of a Web application in which model it has been applied the concept of “Games with a purpose”, and “Input Agreement” has been used to generate text descriptions automatically from audiovisual material. The main difference with other multimedia tagging frameworks is that the game collects tags associated to timestamps with a fraction of the cost that this operation would cost if hired professionals would do it. As a result of the testing, it can be observed that the annotated tags, being evaluated for a human, are relevant to the video context. This could be used for proving the effectiveness of the paradigm called “Human Computation”. A last important contribution of this work refers to the fact that the model developed can be used to generate tags in any language. In particular, the generated test, produced tags for videos in Spanish.