What is Digital Watermarking?
Digital watermarking is a technique of embedding additional information within a digital content such as images, texts, software codes, audio or video signals, etc. The additional information could be anything like image logo, text message, raw watermark bits, etc. By watermarking a content, one gets the idea about the source or owner to which the data belongs.
Basically, in watermarking, one hides a specific message within the digital content. This technique is somewhat similar to steganography as in both of them specific message is hidden within the digital data. However, their individual goal separates the two.
In simplest terms, digital watermarking is a transparent, visible or invisible information patten which is added to any suitable data source by using computer algorithm.
Fundamentals of Digital Watermark
Watermarking is a traditional method that had been in use since past several years. Its one of the traditional and popular example is using invisible ink. Earlier people wrote secret information within a document by making use of invisible ink so that it gets secured from prying eyes.
Initially when digital watermarking came into existence then it was written as two separate words i.e., water mark. Watermark is the information that is to be hidden however, the term itself is indicating that the information which is hidden is as transparent as water.
Digital watermarking enables us to trace the copyright infringement that occurs in social media as it holds the information about the genuine owner of the content.
Let us have a brief idea about some terminologies that are commonly used in the field of watermarking.
- Cover Media: This is the media that can carry the watermark. Also known as original media or host media.
- Watermarked Data: The media that holds the watermark.
- Embedding: The way of adding watermark into the cover media.
- Extraction: The method that extracts the embedded watermark from watermarked data.
- Detection: The way of identifying whether the content contains watermark or not.
- Noise: The disturbance that adds with the watermarked data at the time of transmission.
A noteworthy point over here is that a watermark must be such that it must be detectable from the watermarked content even if the content undergoes intentional or unintentional manipulations. Hence the complete digital watermarking is combination of two processes mainly watermark embedding and watermark extraction.
Concept of Digital Watermarking
As the sharing of information over the internet has become a usual practice hence digital watermarking is in great demand today. When we share files over the internet then the owner is unaware of the fact that any third party is using its content without consent.
The transparent nature of watermark helps in such a way that the cover file and watermarked file are same as watermarks do not change the perception of original file.
Let us now understand the concept of watermarking through the pictorial representation given below:
Whenever one wants to create a watermarked cover WC then watermark data W is to be inserted into the cover object OC. The distortion introduced within the channel can be obtained as the difference between OC and WC.
Types of Digital Watermarking
Watermarking is classified in several ways according to different points of view. Let us understand each separately.
Visible and Invisible: This classification is based on the point of view that whether one can see the embedded watermark through naked eyes or not. Visible watermarks are the ones that can be seen through naked eyes while invisible watermarks are embedded in a way that they cannot be seen by naked eyes.
Images of visible watermark
Robust, Semi-Fragile and Fragile: This categorization of watermark is based on level of resistance of content towards natural noise or artificial modification like attacks.
A robust technique is the one which can detect the hidden watermark within the watermarked data even in the presence of noise or attack. On the contrary, if the technique is unable to detect the presence of noise in the watermarked content then it is o fragile nature.In between there is another category called semi-fragile which shows strong resistance to some noise but weak resistance to some other kind of noise in the watermarked data.
Blind and Non-blind: This classification of watermarking depends on the variation between the original cover file (non-watermarked data) and the watermarked data in order to get the actual watermark. In blind watermarking the extraction or detection does not require the original data. As against, non-blind technique involves the use of original data when extraction at the receiver takes place.
Blind watermarking is more suitable for practical applications than non-blind watermarking.
Public and Private: The public and private classification of watermarking depends on who can recover the original content. When only the authorized users can recover the original host data from the watermarked content then it is known as private watermarking. While in the case when anyone can read the inserted watermark in the host data then it is known as public watermarking.
It is believed by the researchers that private watermarking exhibit superior robustness than public watermarking methodology.
Symmetric and Asymmetric: This categorization of watermarking is based on the parameters used for extraction procedure. When the extraction procedure utilizes the same set of parameters that are used at the time of embedding then it is regarded as symmetric watermarking. The parameters involve secret keys along with information like embedding position and process.
On the other side, an asymmetric technique is opposite to the symmetric one as it uses different set of parameters and keys for embedding and extraction operations.
Asymmetric watermarking is considered to be better than symmetric watermarking.
Reversible and Non-reversible: When the embedded watermark is removed from the watermarked signal and one gets the non-watermarked signal then that watermarking algorithm is of reversible nature. However, such algorithms that cannot transform watermarked signal into non-watermarked signal then it is of non-reversible nature.
Applications
Now, a days, digital watermarking has become a huge part of our lives and so in involves various applications that are as follows:
Content Protection: Any digital media that is embedded with permissible watermark tag allows multiple users to watch as well as copy it. Watermarking acts as security layer over the digital content that prevents unauthorized access of the content.
Example: Audio files have a feature of mute, show explanatory message, stop copying, stop playback an unauthorized digital content.
Content Identification and Management: By watermarking one can insert unique digital identity over the content and this will not cause damage to the actual data so it will not cause interference in user’s enjoyment. Not only this, wide range of digital devices such as computers, video players can manage the digital content.
Example: A watermark that is detected as background program helps to trigger any predefined action like linking to another website.
Content Filtering: At the time of content identification, filtration of digital content can also be performed.
Example: Personal activity can be introduced like pressing a button for more information while watching a video. Also, some advertisement, commercial information appears within the video. Sometimes, content filtering is applied to block certain kind of content to restrict it from watching by the children under the Internet content supervision.
Forensic and Piracy Deterrence: In the field of forensics, a situational metadata like IP address is embedded into the digital content at one or each stage of distribution so that one can retrieve the forensic watermark as evidence.
Hence, one can conclude that digital watermarking is quite a useful technique that is in wide usage in today’s time.
