HPFIT  High Precision Fitting (ESPRIT HPCN PST Activity 24875). Implementation of an industrial robot production system, which is guided in real-time by vision. This was achieved by integrating four cameras and high performance data processing into the closed-loop control interfacing a standard industrial robot. An image-based control scheme was employed.

Depal Implementation of a system for automatic unloading of  piled boxes and box-like objects via an industrial robot. A laser sensor was used to acquire 3D data and model online the objects to be picked. Modelling of objects of variable shape such as sacks (for grasping purposes) has been successfully done.

Hygeiorobot (GSRT-PENED-2000-ED623)- Development and imple-mentation of algorithms for manipulation and navigation control of a mobile robotic system. The final goal of the system was to evaluate the potential of intelligent robot technology in performing assistive tasks within a hospital environment.

SCOVIS (IST-216465)  Self configurable Cognitive Video Supervision using a network of cameras. Work includes tracking, automated configuration of cameras, behavior recognition and semi supervised learning as well as relevance feedback techniques. The system is applied in the automotive industry to automatically learn manufacturing processes involving humans.

Polymnia (IST-2-004357) Development of a novel vision-based system for the automated production of personalised video souvenirs for visitors in leisure and cultural heritage venues. Visitors were identified and tracked through a camera network. The system produces a personalized DVD souvenir allowing visitors to relive their experiences. We identified and tracked visitors by fusing facial and body features.

SemVeillance (GSRT-EHG91) A network of cameras for the identification of security-related events in video content in real time and offline through human detection and tracking. A homography based framework is used for spatial modeling and short term as well as long term behaviors were recognised through machine learning techniques.

Supervision at the production line and semantics extraction (GSRT PENED 03ÅD238). Online detection of semantically meaningful events on the production line, where manual tasks are executed. For that purpose machine learning techniques such as Hidden Markov Models and Support Vector Machines were used.

BOEMIE (Bootstrapping Ontology Evolution with Multimedia Information Extraction)  (IST-FP6-027538).  BOEMIE will pave the way towards automation of the process of knowledge acquisition from multimedia content, by introducing the notion of evolving multimedia ontologies. The results of the analysis from images, video, video OCR, text as well as audio are then probabilistically fused to extract the semantic content and populate the ontologies.

Automated detection of harmful content for sensitive social groups in video data (GSRT PENED 03ED 698). Obscene and violent  is detected and filtered using visual and audio input and fusion using ontological approach.

ITEA-DS  (Intelligent Tools for Emergency Applications & Decision Support) (IST-1999-20254) Development and implementation, at functional prototype level, of a platform of interoperable & tightly integrated tools for Data Acquisition, Situation Assessment (including Risk Analysis), Decision Support and Procedures Management and Reporting for ships.

Marides (Maritime Decision Support) (IST-1999-21129). Goal is the decision making of the chartering departments of the shipping companies and the application and enforcement of the desired business strategy on the fleet of a shipping company by establishing a comprehensive, accurate and real-time on-board chartering expert system.