In this paper, we proposed a novel idea of orientation-free visual secret sharing (VSS). In traditional VSS, before decoding the secret image, users must locate the transparencies in correct direction first. If the
transparency is noise-like, the users will waste much time for checking the correct orientation. In the proposed method, no matter what the stacking orientation is, the secret image will be revealed. Moreover, each generated transparency is non-expansible; thus, its size is the same as the size of secret image. The idea of the proposed method can be also applied to visual secret sharing for multiple secrets.