Secure spread spectrum watermarking for images, audio and video

Abstract

We describe a digital watermarking method for use in audio, image, video and multimedia data. We argue that a watermark must be placed in perceptually significant components of a signal if it is to be robust to common signal distortions and malicious attack. However, it is well known that modification of these components can lead to perceptual degradation of the signal. To avoid this, we propose to insert a watermark into the spectral components of the data using techniques analogous to spread sprectrum communications, hiding a narrow band signal in a wideband channel that is the data. The watermark is difficult for an attacker to remove, even when several individuals conspire together with independently watermarked copies of the data. It is also robust to common signal and geometric distortions such as digital-to-analog and analog-to-digital conversion, resampling, quantization, dithering, compression, rotation, translation, cropping and scaling. The same digital watermarking algorithm can be applied to all three media under consideration with only minor modifications, making it especially appropriate for multimedia products. Retrieval of the watermark unambiguously identifies the owner, and the watermark can be constructed to make counterfeiting almost impossible. We present experimental results to support these claims. A longer version of this document is available at: http://www.neci.nj.nec.com/tr/neci-tr-95-10.ps.

Rotation, scale and translation invariant spread spectrum digital image watermarking

“…In many cases, image contrast may be distorted, for example by a scanner, in which case the e ects of change of contrast must be compensated for in some way. Cox et al [5,6] describe a method known as dynamic histogram warping [7] to carry this out. …”

“…Note that the smaller is the number of bits of core information or \payload" contained in a watermark, the greater the chance of it being communicated without error. Cox et al [7,6] recover a watermark by explicitly computing the correlation between the (noise corrupted) watermark recovered from the image with the perfect watermarks stored in a database. This is a very robust technique for watermark recovery but it is not very useful in practice because of the need for access to the database of marks and the large amount of computation required.…”

“…Since then there has been an increasing use of spread spectrum communications in digital watermarking. It has several advantageous features such as cryptographic security [29,30,6], and is capable of achieving error free transmission of the watermark near or at the limits set by Shannon's noisy channel coding theorem [16,25]. Spread spectrum is an example of a symmetric key [24] cryptosystem.…”

“…Objective criteria for measuring the degree to which an image component is signi cant in watermarking have gradually evolved from being based purely on energy content [16,5,6] to statistical [20] and psychovisual [27,10] criteria. Digital watermarking is also fundamentally a problem in digital communications [16,25,5,6]. In parallel with the increasing sophistication in modelling and exploiting the properties of the human visual system, there has been a corresponding development in communication techniques.…”

“…Information has been embedded using the Discrete Cosine Transform (DCT) [16,34,5,6] Discrete Fourier Transform magnitude and phase [15], Wavelets [16], Linear Predictive Coding [13] and Fractals [9,22]. The key to making watermarks robust has been the recognition that in order for a watermark to be robust it must be embedded in the perceptually signi cant components of the image [16,5,6]. The term \perceptually signi cant" is somewhat subjective but it suggests that a good watermark is one which takes account of the behaviour of human visual system.…”