Content fingerprint is a technique for extracting a set of geometric invariant features from the content to create an image hash that is tolerant to incidental manipulation such as data compression operation while remain sensitive to malicious tampering such as object deletion and addition. The main difference between content fingerprinting and tradition image hash is that later is sensitive to any changes no matter how minute the changes are.
Content fingerprinting finds its applications in multimedia content protection. It is an effective way of combating multimedia piracy. To avoid detection, pirated content is usually manipulated in some way (e.g., slight rotation, recompression, scaling, etc) before distribution. This makes content fingerprinting a superior approach to content protection.