multi-scale curvatures based perceptual vector data hashing techinique for vector content authentication

摘要

The present invention relates to a method of hashing a multi-scale curvature-based visual vector model. A first aspect of the present invention is to generate a multidimensional feature coefficient matrix based on a multi-curvature activity energy distribution of Radius curvature, Turning angle curvature, and Gaussian curvature for all polyline and polygon objects, A first step of obtaining a multi-dimensional intermediate hash coefficient matrix by random mapping by Bell Polynomials; And a second binary hash matrix is obtained by Lloyd-Max quantization on a real-number-type intermediate hash coefficient matrix, and a multidimensional binary hash matrix is divided into a second and a third binary hash matrix by a scale and a curvature type, step; Scale curvature-based visual model model hashing method according to the present invention. The second aspect of the present invention also relates to a multi-scale and curvature-based hash function A hash generation process; And hash extraction function A vector model authentication process; Scale curvature-based visibility vector model hashing method according to the present invention. Thereby, the polylines are grouped on the main layer in the vector data model, the group coefficients are generated according to the first and second curved curvature distributions of the polylines, After generating the final binary hash by the binarization process, the generated hash satisfies the robustness against various attacks and security and uniqueness by the random key. In addition, the proposed method against object attacks such as object simplification, deletion, copying, and hanging has a low error detection probability of about 2 × 10 ^-5 ~ 2.8 × 10 ^-2 compared with the conventional method, The method proposed by the hash uniqueness evaluation of the three types provides an uniquely high probability of 0.014 as compared with the conventional method. In addition, the method proposed in the evaluation of differential entropy based security provides an effect that the entropy of 0.875 ~ 2.149 can be higher than the existing method.