Abstract In this paper we describe our participation in the interactive surveillance event detection task at TRECVid 2012. The system we developed was comprised of individual classifiers brought together behind a simple video search interface that enabled users to select relevant segments based on down~sampled animated gifs. Two types of user -- `experts' and `end users' -- performed the evaluations. Due to time constraints we focussed on three events -- ObjectPut, PersonRuns and Pointing -- and two of the five available cameras (1 and 3). Results from the interactive runs as well as discussion of the performance of the underlying retrospective classifiers are presented.

Url http://doras.dcu.ie/17859/1/dcu-savasa.pdf

Collections Ireland -> Dublin City University -> Publication Type = Conference or Workshop Item
Ireland -> Dublin City University -> Subject = Computer Science
Ireland -> Dublin City University -> Status = Published
Ireland -> Dublin City University -> DCU Faculties and Centres = Research Initiatives and Centres: CLARITY: The Centre for Sensor Web Technologies
Ireland -> Dublin City University -> Subject = Computer Science: Multimedia systems
Ireland -> Dublin City University -> Subject = Computer Science: Information retrieval
Ireland -> Dublin City University -> Subject = Computer Science: Digital video
Ireland -> Dublin City University -> Subject = Computer Science: Machine learning
Ireland -> Dublin City University -> DCU Faculties and Centres = Research Initiatives and Centres