基于极谐变换的鲁棒图像哈希算法

Robust Image Hashing Based on Polar Harmonic Transform

针对现有图像哈希算法普遍存在抗几何攻击能力不强的缺点,提出了一种基于极谐变换(Polar Harmonic Transform,PHT)的鲁棒图像哈希算法。极谐变换是新近发展的一种正交矩变换方法,图像的变换幅值对旋转缩放具有不变性,利用这一特点用图像的极谐变换幅值生成图像摘要。首先对图像进行极谐变换,然后将得到的幅值系数进行筛选量化得到二值序列,最后根据预先设计的密钥对二值序列进行置乱生成图像摘要。仿真结果表明,上述算法对几何攻击和常规信号处理攻击均具有很强的鲁棒性,优于对比算法性能,能够用于图像识别、搜索、认证等方面的应用。

Most existing image hashing algorithms are not robust enough to geometric attacks. To solve this problem, a robust image hashing algorithm based on Polar Harmonic Transform （PHT） was proposed in this paper. The PHT is a recently developed orthogonal moment method, and the magnitudes of PHTs are invariant to image rotation and scaling. According to the feature, the magnitudes of PHTs were quantized into image digest. The images were transformed with Polar Harmonic firstly, then the derived magnitudes were selected and quantized into binary sequence. At last, according to the previous key, the sequence was disturbed to receive the image digest. Simulation results show that the proposed hashing is robust to geometric attacks and conventional signal processing attacks, which outperforms the typical comparative algorithms and can be used in image identification, search and authentication, etc.