Mediahedge’s technical core is formed by the Philips Video Fingerprinting and Database Matching technology. This state of the-art technology enables customers to uniquely identify video items by comparing electronic Fingerprints derived from video segments, and comparing them with fingerprints taken from the original content library.

Even if the video quality is severely degraded due to cropping, scaling, or compression, it is still possible to match the derived fingerprint with the original content fingerprint and to positively identify the content item. This robustness is due to the unique capability of being able to cope with non-exact matches.

The Video Fingerprinting technology is implemented in software, running on standard computing hardware and is designed with a proven server-client-master-slave architecture for flexibility, scalability and reliability.

A video segment is in fact just a few seconds of video, whilst a video item can be anything between one segment and a full length feature film or television event. One digital Fingerprint is taken from each video segment so the Video Fingerprinting technology granularity is superb and even frame-accurate. Furthermore, since the Video Fingerprints are compact, the solution has a high efficiency and performance. As a result the network traffic caused by matching requests is also reduced to a minimum.
 
Mediahedge’s video fingerprinting features at a glance

• Can detect clips as short as 4.6 seconds
• Robust against audio dubbing etc.
• Transparent for multi-lingual content
• No pre-processing required, can identify legacy content
• Proven and scalable system architecture
• Fast and reliable Fingerprint matching
• Robust against severe quality degradation
• Able to handle non-exact matches