分野別セミナー

Functional encryption provides more sophisticated and flexible expression between the encryption key $ek$ and decryption key $dk$ by deriving from attribute vectors $\overrightarrow{x}$ and policy vector $\overrightarrow{v}$, respectively. There is a function $f({\overrightarrow{x}},\overrightarrow{v})$ that determines what type of a user with a secret key $dk$ can decrypt the ciphertext encrypted under $ek$. This allows an encryptor to specify a functional formula as a decryptable policy describing what users can learn from the ciphertext without knowing the decryptor's identities or public keys.
In this talk, we give the explore two geometric-area-based key generation and functional encryption schemes (GeoEnc), where secret keys are associated with a point on a planar coordinate system and encrypt policies are associated with a line (\textit{GeoEncLine} scheme) or a convex polygon ($\textit{GeoEncHull} $scheme). If the attribute point lies on the line or inside the convex hull, the decryption key holder can decrypt the ciphertext associated with the geometric policy such as the line or the convex polygon. The proposed schemes have {policy hiding} as well as {payload hiding} characteristics. We give an evaluation of key distribution in a practical coordinate system and also give a security analysis with a hybrid model. The proposed schemes have many applications as sources for keys generation and policies encryption such as computer graphics security, network topology protection, secure routing and mobile networking, secure multiparty computation, secure GPS/GIS, military area protection, etc.