Abstract Due to tremendous growth in multimedia applications and services, people can easily create, distribute, broadcast and store information. The fact that multimedia content can be easily copied and tampered has motivated a large number of researchers to devise image verification techniques using perceptual image hashing. In this paper, a perceptual image hashing technique is proposed by using difference of Laplacian pyramids. A Laplacian pyramid is a multi-scale representation of an image and can be used to extract stable features. In the proposed scheme, two different pyramids are generated by using filters of different diameters. The difference of Laplacian is calculated to get a unique and robust hash. A number of experiments have been carried out to gauge the effectiveness of the proposed scheme. The results reveal that the proposed technique is robust against non-malicious manipulations and can detect minute level tampering.

中文翻译：

---

使用拉普拉斯金字塔的鲁棒图像哈希方案

摘要 由于多媒体应用和服务的巨大增长，人们可以轻松地创建、分发、广播和存储信息。多媒体内容很容易被复制和篡改，这一事实促使大量研究人员设计使用感知图像散列的图像验证技术。在本文中，利用拉普拉斯金字塔的差异提出了一种感知图像哈希技术。拉普拉斯金字塔是图像的多尺度表示，可用于提取稳定的特征。在所提出的方案中，通过使用不同直径的过滤器生成两个不同的金字塔。计算拉普拉斯算子的差异以获得唯一且稳健的哈希。已经进行了许多实验来衡量所提出方案的有效性。