Forensic watermarks embedded through video codecs help protect OTT content against piracy

Due to the increasing demand for OTT and streaming video content, the adoption of watermarking techniques has also seen a rapid boom. Watermarking approaches help content owners achieve digital authentication as well as copyright protection for their video assets.

While digital videos can be defined as a sequence of still images, video watermarking techniques are very different from image watermarking because of video coding technologies, redundancy between video frames, and attacks specific to video content. Based on the embedding strategy, video watermarking algorithms can be categorized into uncompressed or original video watermarking, codec-based watermarking, and compressed video watermarking.

Codecs are essentially compression technologies and have two components, an encoder to compress the files and a decoder to decompress. There are several types of video codecs available, such as H.261, H.263, ITU-T, ISO/IEC of MPEG1, MPEG-2, MPEG-4, AVC, HEVC, etc. HEVC has evolved as the dominant codec for watermarking applications due to scalable video coding (SVC) facilities that can partially decode a compressed video sequence to a combination of resolution, quality, and frame rate and optimize it on the platform and transmission bandwidth of the user device.

Codec-based video watermarking approaches are popular in safeguarding the interest of copyright holder in DRM protected content on OTT platforms. These approaches primarily rely on the coding structure and generally integrate the watermarking and video-compression processes to reduce their overall complexity. Thus, the watermarking framework should be optimally implemented at the encoder to ensure proper integration. In this approach, the watermark is embedded during video encoding and block based transform coding along with compensation prediction coding are the key aspects.

The principles of video data compression, like quantization and entropy coding, motion estimation, motion compensation, and transformation to spatial redundancy, are used in this process. Since the watermark is embedded in the transform domain coefficient [discrete cosine transform (DCT), discrete fourier transform (DFT), and discrete wavelet transform (DWT)] combined with the encoding process, it does not increase the bitrate of the video. This also allows watermarking algorithms to be created for multiple attacks.

Furthermore, watermark embedding and extraction can be processed simultaneously in real time. However, there is a drawback that accumulation of GOP (group of pictures) can occur while using this approach, since the encoder and decoder need to be modified and the codec might not be able to perform watermark embedding and detection.