|
This decade has witnessed a
major upsurge in terrorist activity across the world. Civil security research
has particular importance for ensuring the safety of the world's citizens at a
time when terrorist threats are more diverse and less predictable. There are
widespread expectations that technology will provide an effective defence
against terrorism – particularly through the use of CCTV. However, despite the
extraordinary progress in real-time processing of CCTV image streams, there are
a number of significant methodological challenges which must be met if automatic
interpretation systems are to be effective. These challenges include segmenting
and tracking individuals in crowded scenes; the ability to characterize threat;
; reducing the number of false alarms; automatic learning of human-oriented
scene structure; and semantic linkage between networks of cameras and other
sensors. To address these challenges, contributions are particular welcome in
the following areas: |
