Tampering estimation watermarking algorithm is proposed based on lifting wavelet and chaotic sequence. The algorithm decomposes the original image into L level integer wavelet based on lifting method, embed spatial location watermark into the first detail coefficients, and embed the estimation watermark into other selected subbands in order to give estimation to the possible attacks. The prominent value of wavelet coefficient, position information of an embedding region and image protection key are used to construct the initial value for the chaotic model, which is sensitive to its initial value and dynamically produces the watermark. At receiver side, the detector firstly decomposes the suspicious image, locates the spatial tampering in the first level subband, and estimates the type of attack operation. Experiments show that this algorithm not only can localize the spatial tampering precisely, but also can estimate the kind of frequency operations, such as compression and filtering
[1]
Hong-jun Zhang,et al.
Bifurcation and chaos in an optically bistable liquid-crystal device
,
1986
.
[2]
Nasir D. Memon,et al.
Counterfeiting attacks on oblivious block-wise independent invisible watermarking schemes
,
2000,
IEEE Trans. Image Process..
[3]
G. C. Langelaar,et al.
Watermarking digital image and video data
,
2000
.
[4]
Deepa Kundur,et al.
Towards a telltale watermarking technique for tamper-proofing
,
1998,
Proceedings 1998 International Conference on Image Processing. ICIP98 (Cat. No.98CB36269).
[5]
A. Murat Tekalp,et al.
Hierarchical watermarking for secure image authentication with localization
,
2002,
IEEE Trans. Image Process..