A watermark is essentially a digital code containing information of the origin, ownership, destination, copy control and transactions that is inserted into a video sequence to provide additional security to DRM protected content.
An effective watermark should typically have a certain set of characteristics which include fidelity (the watermark should be difficult to notice and should not hamper the quality of content), robustness (it should be able to survive common distortions), tamper resistance (it should resist malicious attacks), ease of computation (easy to insert and detect, scalability and affordability), and the ability to coexist with other watermarks. In addition, an efficient watermarking solution should also have a low false positive rate, i.e., an unwatermarked video should not be identified as having a watermark.
The amount of instructions that a watermark can hold is also restricted and the information gets affected if the watermark holds more data. The capacity is also limited by the size of a video sequence. Hence, the data payload must also be assessed while implementing video watermarking solutions. The relative importance of these characteristics depends on the application.
In addition to the aforementioned essential characteristics, in order to be effective as a real-time technique, video watermarking solutions should also have the following additional requisites for compressed video data transmissions:
- Oblivious: It should be viable to extract the watermark information even in the absence of the original unwatermarked data (for example, in case of a set-top-box or recorder which lack the original data).
- Low Complexity: Real-time application and utilization in customer products requires efficiency and speed. Hence, the watermarking techniques should be economical and not too intricate.
- Preserve Host Data Size: The original dimensions of the compressed host data must not be augmented with the watermark. At times, transmission of data over a preset bit-rate channel creates problems such as in hardware decoders where the buffers run out of space. If the dimensions of a compressed MPEG-video stream are enhanced, it might also get difficult to synchronise the audio and video. The encryption of the watermark information needs to be appended to the host data during the watermark insertion process. These encryption and insertion techniques use keys which may differ in time. Therefore, the cryptographic protocols must oversee the key-management intricacies and care must be taken while extending, analyzing and verifying the watermark insertion methods.
