The growing increase in the demand for digital video content has also necessitated the adoption of solutions such as watermarking to offer protection from infringement at the users’ end. Digital watermarking solutions have, thus, become a must have to provide additional security to DRM protected content.
Watermarking works by embedding imperceptible information in the digital file. This information can then be extracted in the event of infringement to identify the exact source of leakage. However, this slight modification made to the digital content is generally irreversible, that is the watermarked version of the content is different from the original un-watermarked one. This means that any further usage of the content would happen on a slightly corrupted version, which makes irreversible watermarking undesirable for sensitive applications, such as military investigation, medical diagnosis, space exploration, etc.
This is where reversible (also called invertible or lossless) watermarking techniques come in. They allow the complete extraction of the embedded data along with restoration of the cover work. Reversible video watermarking techniques can be broadly categorized into robust, fragile, and semi-fragile components. Watermarks in robust schemes can survive normal image frame processing operations, while in fragile techniques, the inserted watermark disappears upon any modification to the watermarked content, thereby revealing that the content has been compromised. In case of semi-fragile schemes, the watermark can survive minor unintentional modifications.
Some newly emerging reversible watermarking approaches which are applied to DRM protected content on OTT platforms have been described below:
- Compression-based reversible watermarking: In this technique, the information needed for recovery of the original image is embedded along with the watermark which requires more space. Hence, a part of the cover image is compressed for data embedding.
- Quantization-based reversible watermarking: Reversible techniques based on the quantization technique are generally fragile. Here, a string of bit values are embedded into the original object based on a quantization scheme, the Quantization Index Modulation scheme being the most traditionally used. The modulation information is then calculated and stored in the watermarked object.
- Expansion-based reversible watermarking: This is a fragile scheme which uses the pixel-based DE (difference expansion) technique and offers a high embedding capacity with low computational complexity.
- Histogram-based reversible watermarking: This is a robust watermarking method where the embedding target is replaced by the histogram bin. The embedding techniques in this scheme are generally block-based and can thus resist some operations.
Depending upon the use case, any one or more of these techniques can be applied to video frames, which are used to detect the source of piracy of premium video content. Using one of these forensic watermarking techniques, along with the effective use of multi-DRM SaaS-based model, can protect the rights of copyright holders in the OTT space.