Protect your content notion

rooty · 2009-04-30, 07:15 AM

This is a completely random idea... If i had the time to write the code i would have...

as you all know piracy content... especially among tube sites... has always been a pain in the ass...

if only someone can write the following

server: download and generate database from the freely available content on most content sites(tubes).... hold database containing hashes to several thumbnails to the content it checked already (this can be applied to photos and videos)

client: extract hashes to content snippets.. and submit to server to see if there is any matches.... the server can also keep the past submitted hashes to run against new content added to the database.

so for example we take a video clip... the name and file md5 hash is irrelevant.. as most thieves change the file names and add watermarks... but surely its posible to extract regions in a thumbnail... like (mid-left +20 px) etc etc and with these create a hash based on the smaller thumbnail extracted from the bigger thumbnail... all based on seconds in the video to a proximity.

who ever have the time to experiment on the idea... it should be a long journey to figure out a properly working algorithm... but it would be do-able.

the one that writes this app can make a fair bit of money one could imagine....

anywhoo back to slaving my ass off writing newsletters

2009-04-30, 07:15 AM	#1
rooty A woman is like beer. They look good, they smell good, and you'd step over your own mother just to get one! Join Date: Sep 2008 Posts: 51	Protect your content notion This is a completely random idea... If i had the time to write the code i would have... as you all know piracy content... especially among tube sites... has always been a pain in the ass... if only someone can write the following server: download and generate database from the freely available content on most content sites(tubes).... hold database containing hashes to several thumbnails to the content it checked already (this can be applied to photos and videos) client: extract hashes to content snippets.. and submit to server to see if there is any matches.... the server can also keep the past submitted hashes to run against new content added to the database. so for example we take a video clip... the name and file md5 hash is irrelevant.. as most thieves change the file names and add watermarks... but surely its posible to extract regions in a thumbnail... like (mid-left +20 px) etc etc and with these create a hash based on the smaller thumbnail extracted from the bigger thumbnail... all based on seconds in the video to a proximity. who ever have the time to experiment on the idea... it should be a long journey to figure out a properly working algorithm... but it would be do-able. the one that writes this app can make a fair bit of money one could imagine.... anywhoo back to slaving my ass off writing newsletters __________________ Nothing to see here really - My boring blog