An Event Detection dataset


These pages describe an Event Detection dataset (EDds), a corpus of video sequences created to provide a representative test-set  for comparing event recognition algorithms. Morever, manual annotations of the events of interest are also provided.

The dataset is focused on two types of human-related events: interactions and activities. The former refers to actions performed among several humans or environmental objects (e.g., get objects). The latter considers human actions not involving physical interaction (e.g. walking). In particular, two activities (Hand Up and walking) and three human-object interactions (Leave, Get and Use object) have been annotated. 

 For more details please refer to the "Content" section.

Related publications (using dataset) :

[1] J.C. SanMiguel, M. Escudero-Viñolo, J. M. Martínez and J. Bescós, “Real-time single-view video event recognition in controlled environments”, in Proc. of the IEEE 9th Int. Workshop on Content-Based Multimedia Indexing, CBMI2011, Madrid (Spain), June 2011, pp. 91-96.

Work partially supported by the Spanish Goverment under project TEC2011-25995 (EventVideo).