Histogram-based image hashing for searching content-preserving copies

  • Authors:
  • Shijun Xiang;Hyoung Joong Kim

  • Affiliations:
  • Department of Electronic Engineering, School of Information Science and Technology, Jinan University, Guangzhou and Institute of Software, Chinese Academy of Sciences, Beijing, China;Graduate School of Information Management and Security, Korea University, Seoul, Korea

  • Venue:
  • Transactions on data hiding and multimedia security VI
  • Year:
  • 2011

Quantified Score

Hi-index 0.00

Visualization

Abstract

Image hashing as a compact abstract can be used for content search. Towards this end, a desired image hashing function should be resistant to those content-preserving manipulations (including additive-noise like processing and geometric deformation operations). Most countermeasures proposed in the literature usually focus on the problem of additive noises and global affine transform operations, but few are resistant to recently reported random bending attacks (RBAs). In this paper, we address an efficient and effective image hashing algorithm by using the resistance of two statistical features (image histogram in shape and mean value) for those challenging geometric deformations. Since the features are extracted from Gaussian-filtered images, the hash is also robust to common additive noise-like operations (e.g., lossy compression, low-pass filtering). The hash uniqueness is satisfactory for different sources of images.With a large number of real-world images, we construct a hash-based image search system to show that the hash function can be used for searching content-preserving copies from the same source.