Journal of Applied Sciences ›› 2016, Vol. 34 ›› Issue (6): 691-701.doi: 10.3969/j.issn.0255-8297.2016.06.005

• Signal and Information Processing • Previous Articles     Next Articles

Robust Image Hashing Based on Content Structure Diagram

LI Xin-wei1,2, XIA Xiu-zhen1   

  1. 1. School of Electrical Engineering and Automation, Henan Polytechnic University, Jiaozuo 454000, Henan Province, China;
    2. Key Laboratory of Control Engineering of Henan Province, Jiaozuo 454000, Henan Province, China
  • Received:2016-04-26 Revised:2016-07-07 Online:2016-11-30 Published:2016-11-30

Abstract:

Robust image hashing is proposed for improving robustness and matching efficiency of copy detection system. Gabor transform coefficients of the image are used to construct structure diagrams. They are transformed from the Cartesian coordinates to polar coordinates, and normalized to obtain new diagrams. The weighted sum of their subblock pixels are calculated to obtain feature vectors, which are then quantized to produce the final binary hash codes. The structure diagrams based on Gabor coefficients proposed in this work are robust and distinctive. The distorted hash code fusion and double keys used in quantization further improve robustness, distinctiveness and compactness of the algorithm. The proposed method is compared with several representative image hashing approaches such as the non-negative matrix factorization hashing, radial and angular shape context hashing, ring partition and invariant vector distance hashing using a large image database. The results show that the overall performance of the proposed algorithm is significantly better than the other methods in terms of precision and recall rates, as well as matching efficiency.

Key words: robustness, hash code fusion, image hashing, content structure diagram, Gabor transform

CLC Number: