An annotated video database for abandoned-object detection in a cluttered environment

Abstract

A video database is described for testing video surveillance systems for the automatic detection of abandoned objects in a cluttered environment. The database requirements and development are included along with a full description of its contents. The complete database comprises a total 6 multi-object, 56 single-object and 4 no-object (for reference purposes) footages, acquired with two different cameras and two different light conditions, yielding an approximate total of 8.2 hours of video. Database annotation was performed with a specialized tool, developed in the context of this particular work, that provides a bounding box for each object (or its several parts in case of partial occlusion) in every video frame. The complete dataset constitutes a valuable tool for designing and testing automatic video-based systems for abandoned-object detection.