Image hashing algorithm based on multidimensional features and spatial structure

To improve the comprehensiveness of hash algorithms, an image hash combining multidimensional features and spatial structure is proposed. Firstly, the multidimensional feature matrix is constructed using the quaternion model, and its dimensionality is reduced by quaternion singular value decomposition to obtain the multidimensional image features; at the same time, the spatial structure of the image is extracted from the local projection transformation information and global projection difference information to obtain the spatial structure features. The results demonstrate that the scheme has high robustness and distinguishability, and has excellent capability in copy detection.