Robust Image Hashing Based on Selective

摘要: Robust image hashing, which maps the perceptual contents of to a short digest, is key technique for tackling challenges content-based indexing, searching and copyright protection.In this letter, wepropose quaternion invariance based hashing algorithm that can fuse complementary visua lf eatures compact hash. The proposed leverages polar cosine transform holistically capture spatial chromatic characteristicsofdigitalimage,androtation-invariantfeaturesare derived from phase information in frequency domain. In addition, we also propose an theoretic metric feature quality assessment ag reedy strategy select optimal subset features hash computation. Extensive experiments over large database demonstrate work shows higher content identification accuracy than most competing algorithms, resulting easy compute.