We propose a quantum secret sharing scheme between $m$-party and $n$-party using three conjugate bases, i.e. six states. A sequence of single photons, each of which is prepared in one of the six states, is used directly to encode classical information in the quantum secret sharing process. In this scheme, each of all $m$ members in group 1 choose randomly their own secret key individually and independently, and then directly encode their respective secret information on the states of single photons via unitary operations, then the last one (the $m$th member of group 1) sends $1/n$ of the resulting qubits to each of group 2. By measuring their respective qubits, all members in group 2 share the secret information shared by all members in group 1. The secret message shared by group 1 and group 2 in such a way that neither subset of each group nor the union of a subset of group 1 and a subset of group 2 can extract the secret message, but each whole group (all the members of each group) can. The scheme is asymptotically 100% in efficiency. It makes the Trojan horse attack with a multi-photon signal, the fake-signal attack with EPR pairs, the attack with single photons, and the attack with invisible photons to be nullification. We show that it is secure and has an advantage over the one based on two conjugate bases. We also give the upper bounds of the average success probabilities for dishonest agent eavesdropping encryption using the fake-signal attack with any two-particle entangled states. This protocol is feasible with present-day technique.
Read full abstract