New means for real-time access of visual material
Content-based-retrieval systems have been emerging in recent years as a more powerful and efficient alternative to the World Wide Web for Internet content. The search principles are based upon measures of similarity whereby the queries match examples or prototypes to known media types. The end result is a flexible feature based retrieval engine, which can be trained on user requirements and displays interesting characteristics. Prerequisites for a flexible engine are high image quality and an expert classification of the training set. This is performed with image indexation, where mobile agents transport the content-based-retrieval vector extraction and collection algorithm to the servers of the image providers. As such, this system presents a service that permits mobile agents to communicate the results to the user through ordinary browsers. The principle benefits for current products enables this search engine to explore information spaces and judge their information relevance by their position and colour density within a sphere. It can also provide interactive relevance feedback and interactive clustering of the results sets. Therefore this software is a flexible and customisable retrieval engine that can perform advanced image analysis and indexing, and content-based video/image retrieval. Further, this search engine has the ability to enable the user to visually judge and manipulate their search results, which is of exceptional value to media broadcasting companies. Also the inbuilt security systems which are crucial for both agents and hosts, means that this search engine will benefit organisations dealing with national and international security. Because the field of content-based-retrieval is continually evolving into a more robust and applicable mechanism, this software can be easily installed and removed if necessary. An additional bonus to this particular system is because it is generic and thus can be used in any image retrieval system. The German institute behind this new and advanced retrieval system is currently looking towards interested partners with a view for information exchange and or shared training.
