Fingerprinting

When you're on the lookout for revenue-damaging piracy, speed and accuracy are non-negotiable.

There is a lot of video content available, whether via the web or digital devices and applications. So when a content monitoring service is searching for instances of illegally shared content, it's essential to be able to quickly and correctly match suspicious material against the original.

FMTS' advanced video fingerprinting technology underpins our world class monitoring services. Fully automated and highly scalable, our Emmy® award-winning fingerprinting technology matches content against the original with unrivalled speed and accuracy.

The automation behind FMTS fingerprinting enables monitoring at huge scale. Our Monitoring services locate, capture and analyse on average over 26,000 hours – or more than 3 years – of video every day. 

How fingerprinting works

Video fingerprinting is a matching process, just like human fingerprinting. The first step is to generate a video fingerprint based on the original video.

This is video that FMTS is tasked with protecting and which is provided by our customers. Just like human fingerprints are unique to an individual, video fingerprints are also unique. They help to automatically identify a video amongst millions of other videos with forensic precision.

Once fingerprints of the original video are generated, they are stored in a fingerprint database. Automation and scale matter: video fingerprints are very lightweight, allowing fingerprints from millions of hours of video to be quickly and efficiently stored and scanned.

The purpose of the fingerprint database is to locate matches against any videos captured by FMTS monitoring systems. Once a video is captured, a fingerprint of it is generated and compared against the fingerprint database. If a match is found within the database then it's confirmed that the video is the same video as the original, and further actions are automatically invoked.

Cloud-native

FMTS deals with the technical complexities regarding the generation of video fingerprints via its cloud-native fingerprint generation services. These support a host of industry-standard video delivery formats for all content types, including live, linear and VOD. FMTS supports RTMP, RTMPS, HLS, DASH, and SRT for live and linear broadcasts and all common video containers and encoding formats for VOD.

Advertising-aware

Advertisements which form part of the original video, for example as part of a live event or linear channel, can cause issues for fingerprinting systems since they do not form part of the actual content you own or wish to protect. Therefore fingerprint matching against advertisements can result in "false positive" reports.

FMTS fingerprinting systems support industry-standard SCTE-35 markers to automatically identify advertisements within the original video stream, suspending fingerprint generation to ensure that they do not form part of the original video content.

Monitoring at scale

Large-scale monitoring is only possible through the implementation of innovative technologies that automate the capture and recognition processes – including video fingerprinting. Whilst it is possible to employ teams of people to manually match videos (as some monitoring services do), this is of course a very small-scale approach. It limits the amount of video that can be processed in real time, massively increases the time taken to analyse each video, and naturally introduces the inevitable factor of human error.

Automation = speed, scale and accuracy.

That's the FMTS difference.

Our technology isn't just advanced – it's industry-leading. But don’t take our word for it. Dive into our case studies and see why the world’s leading brands trust FMTS.