We consider the situation for public-key encryption that the adversary knows the randomness which was used to compute the ciphertext.
In some practical scenarios, there is a possibility that the randomness is revealed.
For example, the randomness used to make a ciphertext may be stored in insecure memory, or the pseudorandom generator may be corrupted.
We first formalize the security notion on this situation as ?the semantic security with the randomness revealed? and
?the non-malleability with the randomness revealed.?
In addition to the formalization, we focus on the scheme, 3-round OAEP, and prove that this scheme satisfies our security notion.