The main objective of SPOTTER is to develop intelligent video surveillance techniques. In particular, it fosters innovative technology for automatically finding the occurrences of objects/regions of interest (e.g., persons, cars, etc) via specialized algorithms embedded directly in the IP video cameras. The genuine novelties of this approach are in the implementation of state of the art, high-level, video processing techniques which will be capable of running in real-time on dedicated surveillance hardware. This involves:
(i) Development of advanced DROP (Distinctive Regions Of interest or Patterns) retrieving algorithms dedicated and optimized for running on low resources embedded hardware platforms (e.g., IP video cameras);
(ii) Development of techniques for real-time identification and tracking capabilities from multiple video sources;
(iii) Development of techniques for algorithms which can dynamically adjust the running parameters according to task variables (generic nature).