Fast Fallback Watermark Detection Using Perceptual Hashes

Abstract

Forensic watermarking is often used to enable the tracing of digital pirates that leak copyright-protected videos. However, existing watermarking methods have a limited robustness and may be vulnerable to targeted attacks. Our previous work proposed a fallback detection method that uses secondary watermarks rather than the primary watermarks embedded by existing methods. However, the previously proposed fallback method is slow and requires access to all watermarked videos. This paper proposes to make the fallback watermark detection method faster using perceptual hashes instead of uncompressed secondary watermark signals. These perceptual hashes can be calculated prior to detection, such that the actual detection process is sped up with a factor of approximately 26,000 to 92,000. In this way, the proposed method tackles the main criticism about practical usability of the slow fallback method. The fast detection comes at the cost of a modest decrease in robustness, although the fast fallback detection method can still outperform the existing primary watermark method. In conclusion, the proposed method enables fast and more robust detection of watermarks that were embedded by existing watermarking methods.

Fast and Blind Detection of Rate-Distortion-Preserving Video Watermarks

“…Note that a recompression with QP 𝑎 = 47 results in a very low-quality video. Watermark detection was performed for each attacked video, using the slow watermark detection [11], the proposed fast watermark detection using NS-DCT-DST hashes, as well as the proposed fast detection method using TIRI-DCT hashes from the state of the art [12]. The FP probability was set to 𝑃 fp = 10 −6 , corresponding to a threshold 𝑇 ≈ −4.7534.…”

“…Hence, this paper proposes to utilize perceptual hashes to perform fast and robust watermark detection. It has been shown that applying perceptual hashes on watermarks has merit as well [12]. However, the robustness of this fallback system decreased due to using the hashes, and the system is still non-blind.…”

Adaptive Video Watermarking against Scaling Attacks Based on Quantization Index Modulation

“…Watermarking can be executed in a spatial domain [4,13,23,28,29] or a transform domain [3,[5][6][7][8][9][10][11][12]14,15,18,19,21,22,[24][25][26][27][30][31][32][33][34][35]. The spatial domain watermarking has high efficiency by modifying the pixels of the host images or frames directly to embed the watermark.…”

“…The spatial domain watermarking has high efficiency by modifying the pixels of the host images or frames directly to embed the watermark. In comparison, the transform domain watermarking modifies the coefficients obtained by transformation, such as the Discrete Cosine Transform (DCT) [8][9][10]18,19,21,22,24,26,30,31], Discrete Wavelet Transform (DWT) [7,11,12,14,[25][26][27]32,33], Singular Value Decomposition (SVD) [6,12,25,26], Discrete Multiwavelet Transformation (DMT) [3], Contourlet Transform [6], Dual-Tree Complex Wavelet Transform (DT CWT) [5,15,34], and so on. They possess better imperceptibility and robustness, but their computational complexity is higher than that of the spatial domain watermarking.…”

Common Gabor Features for Image Watermarking Identification

“…Two metrics are utilized to assess the robustness of our proposal: (1) true-positive rate (TPR), and (2) false-negative Rate (FNR) [34]. Equation (2) calculates true positive (TP), which reflects correctly identified embedded watermarks.…”