2 ms·
I thought that the random collision for md5 was way higher than that. If it's that low, you're right, this would work. I'm not sure I understand the part about
by Zababa 5y ago
I thought that the random collision for md5 was way higher than that. If it's that low, you're right, this would work. I'm not sure I understand the part about the pertubation.
- DSingularity 5y agoIt’s actually 1/(2^128). If the attacker does not know how the image will be perturbed prior to hashing then he cannot generate an image which matches with known CSAM.