The ever-growing amount of information being recorded by surveillance cameras also means spiralling costs for operators, in terms of storage filtering and human resources in data checking. In addition, there are also concerns over the implications for privacy when people's movements in public spaces are being tracked. The EU-funded MOSAIC (Multi-modal situation assessment & analytics platform) project addressed these issues by developing a system embedded with automated detection, recognition, geo-location and mapping. Project partners built surveillance systems by employing decision support technologies that enable cameras to filter out unimportant events, meaning surveillance is more focused and targeted. The solutions are based on video analytics using data ontology methods that search tags and fuse data from both distributed multimedia sensors, sources and information databases. Scientists developed and validated several innovations. These include system architecture to support wide area surveillance with edge and central fusion and decision support capabilities, and algorithms that enable different multimedia information correlation. Yet other innovations are tools and techniques that extract key information from video, uncontrolled text and databases. MOSAIC allows for a more focused and targeted approach to surveillance thanks to the pre-filtering of unimportant events. This substantially reduces network traffic. It minimises false alerts, is cheaper to deploy, operate and maintain, as well as less bandwidth and storage intensive. The technology integrates advanced criminal and social network analysis, together with text and data mining methods. Emphasising the need for further surveillance and monitoring helps users make better decisions. MOSAIC's smarter surveillance solutions will ensure high levels of security at lower cost and result in fewer infringements on the right to privacy thanks to less video footage.
Video analytics, personal privacy, surveillance cameras, MOSAIC, situation assessment