The increasing adoption of video surveillance makes it possible to watch over sensitive areas and identify people responsible for damage, theft and violence. However, when such events are not detected immediately, the subsequent video analysis can be a long and tedious task. The aim of this paper is to present a technique that allows a human investigator to focus only on those parts of a video showing the event as it unfolds, and so helping to save on the time needed to identify and understand how it happened. The presented technique creates a single interactive image of the whole video that shows everything that happened in the scene. The human investigator can then select an area of interest and those parts of the video related to that specific area will start to play. Copyright © 2010 ACM.
Video abstraction and detection of anomalies by tracking movements
BUONO, PAOLO;
2010-01-01
Abstract
The increasing adoption of video surveillance makes it possible to watch over sensitive areas and identify people responsible for damage, theft and violence. However, when such events are not detected immediately, the subsequent video analysis can be a long and tedious task. The aim of this paper is to present a technique that allows a human investigator to focus only on those parts of a video showing the event as it unfolds, and so helping to save on the time needed to identify and understand how it happened. The presented technique creates a single interactive image of the whole video that shows everything that happened in the scene. The human investigator can then select an area of interest and those parts of the video related to that specific area will start to play. Copyright © 2010 ACM.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
