QuOIMA - Open Source Integrated Multimedia Analysis
The automatic monitoring of both traditional, but especially the new social media sources makes it possible to identify early risk indicators and factors during crises and disasters. For this reason, crisis and disaster management need to quickly identify patterns and trends.
Today, crises and disasters are closely and continuously followed by traditional and social media alike. There is not a single minute that they do not document. The information of social media sources (Facebook, Twitter, and others) represent an often untapped potential for crises management centers. By combining these sources with content from TV or newspapers, different or complementary information can be accessed. Due to its non-homogeneous nature this enormous wealth of information, however, can be difficult to overview, channel and utilize.
The systems available in national and international crisis and disaster management centers do not allow analysing all media channels – particularly social media – in an automated way. Specifically, the continual advances in the areas of multimedia and social media, as well as the increasing dynamics of information, require the development of appropriate methods.
The automatic monitoring of both traditional and new social media sources makes it possible to identify early risk indicators and factors for crisis and disaster management. Subsequently patterns and trends can be identified quickly. The resulting comprehensive awareness of the situation enables an early and rapid response to potential crisis situations.
During the course of a crisis relevant (sub-)events are automatically identified from the traditional and social media in the form of text, pictures and video documents. The resulting multimedia documents are clustered according to various criteria (time, location, visual similarity etc.). Hence the information can be presented to the experts in crisis and disaster management in a more structured way.
The research and development of algorithms and methods to achieve the goals listed above are the focus of the QuOIMA project. The automatic analysis of the multi-media contents from open sources helps to come by the wealth of information.
To achieve a high public acceptance for using social media in disaster management it is of highest importance to take care that existing legal frameworks for privacy protection are obeyed. Based on internationally already existing guidelines and the applicable laws, a Privacy Impact Analysis guide was developed. It is worth noting that Austrian data protection law provides a special regulation to enable the immediate use of information in the case of a disaster.