Tuesday, October 12, 2010
Video and Audio Content Identification–Fingerprinting
Video and audio content can be identified without modifying the content by using fingerprinting. During the IBC trade show, Alex Terpstra from Civolution explained to me that there are some key requirements for fingerprinting which depend on the objectives of the content owner or distributor. Some of the requirements include reliability, robustness, and low complexity. Fingerprinting systems must be able to accurately identify the media regardless of the content, platform, or device it is operating on. Having a fingerprinting system that has high robustness allows for the identification of media that has been transformed (such as being recorded by a camcorder in a HD TV screen) where the underlying content may be distorted. Fingerprinting analysis software (clients) may be installed on devices that have limited processing capabilities (such as TV STBs). The fingerprinting system may need to balance between each of these requirements (such as robustness and complexity) to achieve the desired objectives.