Digital Watermarking

Computer Science; Digital Media; Security

Digital watermarking protects shared or distributed intellectual property by placing an additional signal within the file. This signal can be used to inform users of the copyright owner's identity and to authenticate the source of digital data. Digital watermarks may be visible or hidden.

Digital watermarking is a technique that embeds digital media files with a hidden digital code. It was first developed in the late twentieth century. These hidden codes can be used to record copyright data, track copying or alteration of a file, or prevent alteration or unauthorized efforts to copy a copyrighted file. Digital watermarking is therefore commonly used for copyright-protected music, video, and software downloads. Governments and banks also rely on it to ensure that sensitive documents and currency are protected from counterfeiting and fraud.

A paper watermark is an image embedded within another image or a piece of paper. It can be seen by shining light on the image. Watermarks are used on banknotes, passports, and other types of paper documents to verify their authenticity. Similarly, digital watermarking involves embedding data within a digital signal in order to verify the authenticity of the signal or identify its owners. Digital watermarking was invented in the late 1980s or early 1990s. It uses techniques that are also used in steganography (the concealment of messages, files, or other types of data within images, audio, video, or even text).

Most digital watermarks are not detectable without an algorithm that can search for the signal embedded in the carrier signal. In order for a carrier signal to be watermarked, it must be tolerant of noise. Noise-tolerant signals are generally strong signals that resist degradation or unwanted modulation. Typically, digital watermarks are embedded in data by using an algorithm to encode the original signal with a hidden signal. The embedding may be performed using either public- or private-key encryption, depending on the level of security required.

One way to classify digital watermarks is by capacity, which measures how long and complex a watermarking signal is. The simplest type is 1-bit watermarking. This is used to encode a simple message that is meant only to be detected or not (a binary result of 1 or 0). In contrast, multibit watermarking embeds multiple bits of data in the original signal. Multibit systems may be more resistant to attack, as an attacker will not know how or where the watermark has been inserted.

Watermarks may also be classified as either robust or fragile. Robust watermarks resist most types of modification and therefore remain within the signal after any alterations, such as compression or cropping. These watermarks are often used to embed copyright information, as any copies of the file will also carry the watermark. Fragile watermarks cannot be detected if the signal is modified and are therefore used to determine if data has been altered.

In some cases, a digital watermark is designed so that users can easily detect it in the file. For instance, a video watermark may be a visible logo or text hovering onscreen during playback. In most cases, however, digital watermarks are hidden signals that can only be detected using an algorithm to retrieve the watermarking code. Reversible data hiding refers to cases in which the embedding of a watermark can be reversed by an algorithm to recover the original file.

A primary function of digital watermarking is to protect copyrighted digital content. Audio and video players may search for a digital watermark contained in a copyrighted file and only play or copy the file if it contains the watermark. This essentially verifies that the content is legally owned.

Certain types of programs, known colloquially as crippleware, use visible digital watermarks to ensure that they are legally purchased after an initial free evaluation period. Programs used to produce digital media files, such as image- or video-editing software, can often be downloaded for free so users can try them out first. To encourage users to purchase the full program, these trial versions will output images or videos containing a visible watermark. Only when the program has been registered or a product key has been entered will this watermark be removed.

Some creators of digital content use digital watermarking to embed their content with their identity and copyright information. They use robust watermarks so that even altered copies of the file will retain them. This allows a content owner to claim their work even if it has been altered by another user. In some cases, watermarked data can be configured so that any copies can be traced back to individual users or distributors. This function can be useful for tracing illegal distribution of copyrighted material. It can also help investigations into the unauthorized leaking of sensitive or proprietary files.

More recently, digital watermarking has been used to create hidden watermarks on product packaging. This is intended to make it easier for point-of-sale equipment to find and scan tracking codes on a product. Digital watermarking is also increasingly being used alongside or instead of regular watermarking to help prevent the counterfeiting of important identification papers, such as driver's licenses and passports.

—Micah L. Issitt