| Summary: | 碩士 === 國立臺北科技大學 === 資訊工程系所 === 93 === With the advancement of multimedia and networking techniques, the problem of unauthorized use of digital contents becomes more and more serious. Therefore, media hashing becomes an important technique for copyright tracing and protection. In this thesis, a novel robust image hashing scheme is proposed. We extract the identification information of images through the set partitioning in hierarchical trees (SPIHT) algorithm, widely employed in image compression. The significance maps, the spatial distributions of significant wavelet coefficients, are formed from the sorting pass of SPIHT. We build the hash values from the autocorrelograms of the significance maps. To verify the robustness of the proposed methods, experiments are conducted on the Stirmark benchmarking system. The proposed hash sequence achieves good performance with reasonable complexity.
|
|---|
