With the increasing number and variety of camera installations, unsupervised methods that learn typical activities have become popular for anomaly detection. In this article, we consider recent methods based on temporal probabilistic models and improve them in multiple ways. Our contributions

to provide a smooth motion detection map. Then, the obtained change detection map is combined with a geometry module that performs a soft auto-calibration to estimate a measure of congestion of the observed area (platform). Extensive experimental results in a metro station of a metropolitan city demonstrates

;hidden' state for the first 20 frames. Since an actual person blob should be well segmented for a longer period of time, ghosts can often be detected by their instability over time. The setting for our experiments is an actual train station platform on a normal business day. The scenarios recorded for our

Abstract not found

1. Networking trends 2. QoS over data networks 3. Label switching 4. Gigabit, 10 Gb Ethernet, RPR 5. Storage area networks 6. IP over DWDM 7. Wireless 8. Voice over IP

This thesis presents CraneTracker, a multi-modal sensing and communication system for monitoring migratory species at the continental level. By exploiting the robust and extensive cellular infrastructure across the continent, traditional mobile wireless sensor networks can be extended to enable reliable, low-cost monitoring of migratory species. The developed multi-tier architecture yields ecologists with unconventional behavior in-formation not furnished by alternative tracking systems at such a large scale and for a low-cost. The simulation, development and implementation of the CraneTracker soft-ware system is presented. The system is shown effective through multiple proxy deploy-

the condition that they attribute it, that they do not use it for commercial purposes and that they do not alter, transform or build upon it. For any reuse or redistribution, researchers must make clear to others the licence terms of this work.