The technology developed by Advestigo uses different techniques to extract content information from different media types, or Elementary Media - Text, Image, Sound.
Composite documents such as video (audio and image content) or office documents (text and graphics) are first broken down into their different media types, each of which are then analysed for their specific features. Depending on the type of source, different methods may be used to break down the document into its various media types. For example, office documents can be parsed using a formal analysis to identify their constituent elements, whereas video streams require a more complex approach to identify the different elements for analysis.
Video breakdown is based on analysis of static images and audio streams. Static images are built from video segments which correspond to single continuous sequences, automatically detected using signal analysis techniques. This breakdown is independent of the specific encoding method used for the video stream. In a similar manner to the way audio signals are analysed, this fingerprinting of fixed images ensures that copies can be recognised even when local parameters have been altered (changes in background colour, size, format, "pan-and-scan"...). Similarly, cut-and-paste or sampling operations can also be detected.
|
