Robust media search (RMS) is a similarity-based search technology that employs feature data extracted from signal fragments of audio/visual contents. RMS offers robust high-speed and high-precision recognition of targets without using text information such as keywords and metadata. It enables us to retrieve related information only using media data and is suitable for various types of practical applications, such as robust identification of back-ground music titles in noisy environments, high-speed search of huge media databases, and high-precision identification of target objects in practical environments, even when a situation calls for instant reaction.

Please click the thumbnail image to open the full-size PDF file.