Originally Posted by
RavenSword So I'm of the opinion that as long as it's being done for good I'm ok with the scanning. I'm pretty sure it's also that law that if these companies find their users with images like these they are required to report. And from what I heard, maybe I heard incorrectly, but I heard that these types of images have a special id tag or something so it's fairly easy to find them. So maybe there only scanning for those special tags?
They use something called PhotaDNA, MS invented it and Google, Facebook, Twitter amongst others use it.,,,,
PhotoDNA creates a unique signature for each image, similar to a fingerprint, to help pictures be matched. This is done by converting the picture into black-and-white, resizing it and breaking it into a grid. Each grid cell is then analysed to create a histogram describing how the colours change in intensity within it, and the information obtained becomes its "DNA".
The technique means that if a copy of a flagged photo appears in one of Microsoft's user's accounts, the firm can be alerted to the fact without its staff having to look at the picture involved.
Because the amount of data involved in the "DNA" is small, Microsoft can process and compare images relatively quickly.
"[It] allows us to find the needle in the haystack," says promotional material for the software.