Quote:
Originally Posted by brandonstills
For the more technically inclined.
It's nice that they can keep it in O(log n). That leads me to believe that they don't have to compare each fingerprint to every other fingerprint to find a match. They probably store it in a manner in which they specify the video generally and then narrow down more.
That would suggest that they have the ability to find similar videos as well.
|
Quote:
Originally Posted by brandonstills
Seems like this technology requires being able to feed the videos from a site into the engine to fingerprint and then compare. There seems to be a technical and also legal flaw to this.
Technical: You have to write a spider for each site individually because they protect each video. They can detect how many videos you "watch" and if it is over a certain threshold they simply ban you. If they think you are spidering their videos they just have to change the authentication method. IP's suspected of spidering can be banned outright and shared across the network of piraters.
Legal: IANAL but reverse engineering something in order to break its protection is usually considered illegal. The DMCA might have some provisions to allow it or might add some in the future but that still leaves the technical problems mentioned above.
On an unrelated note, this technology has yet further applications. It would be cool for someone to be able to upload a clip or screenshot from a movie and then have the site report back, what movie is this from? Which can ultimately be used for things like, who is this girl?
|
Many thanks - I certainly hadn't considered the legal aspects of reverse engineering. However, I would have thought analysing a publicly available stream was perfect valid... it isn't reversing to find a fingerprint, since there is none there (in my own train of thought).
In any case, thanks for the technical breakdown - that gives me a great idea for something else that I'm stuck on. Really, thanks - it gives me the push I need.