Improving Multi-Histogram-Based Reversible Watermarking Using Optimized Features and Adaptive Clustering Number

Abstract

For the multi-histogram-based reversible watermarking (MHRW) scheme, the performance greatly depends on the multi-histogram construction, which remains a challenge in this field. To generate more desirable multi-histograms, this paper improves the MHRW using the Fuzzy C-Means (FCM) clustering technique by developing the following approaches: 1)optimize the original feature set, 2)adopt an alternative FCM (AFCM) clustering method, and 3)determine adaptively the optimal clustering number for low embedding rates. These approaches are then integrated to bring about the proposed scheme, i.e., the improved MHRW (IMHRW). Extensive simulations show that the proposed scheme improves the performance of multi-histogram-based reversible watermarking, and it is comparable to or even better than the state of the arts. This thus demonstrates the feasibility and effectiveness of the proposed scheme.

Keywords

• Reversible watermarking
• multiple histograms
• prediction error
• fuzzy C-means clustering
• local feature