Multimedia communication creates and plays a vital role in various fields in the society of today's world, which includes economics, politics, militaries, industries as well as entertainment, etc. So it is very much important for securing the data related to multimedia by giving separate identity, integrity, confidentiality, or ownership (Babu et al., 2021). The security related to multimedia deals with the problems mainly related to data encryption, digital watermarking, management of digital rights, as well as, authentication of multimedia. Newly introduced technologies take advantage of the processing as well as communications regarding multimedia security. Multimedia security is a protection system that is based on content. The content consists of various media forms like imagery, video, audio graphics, and texts in a varied range of digital forms. Proper security for any technological system plays a vital role in every aspect. Security saves the data from copyright and misuse. Watermarking has become one of the important methods for the protection of digital elements. Digital watermarking implants a signal in the original element and as a result, the signal can detect the owner easily. But it is easy to destroy fragile watermarks (Giri et al., 2021).
A. Lab results
DAM(Digital Asset Management) comprises tasks based on management and decisions involving ingestion, cataloging, annotation retrieval, and storage as well as digital assets' distribution. The main contents are image, document, video, and sounds, etc. cooperate enterprise manages media content of its own traditionally. Global management enterprise shares content widely. Contents are delivered over various channels through TV, mobiles, the internet, print, etc (Bhat et al., 2021).
DRM( Digital Rights Management) is a group of technologies used by the manufacturers of hardware, copyright holders, publishers, as well as individuals who wish to use digital content as well as devices.
Watermark helps in hiding the information of the data of multimedia. The main applications of digital watermarking are broadcast monitoring, transaction tasking, proof of ownership as well as authentication of content (Honsy et al., 2020).
Microsoft play-ready technology allows the seamless sharing of protected data content between different devices and people. Play ready is the principal DRM system for the protection of content.
Three requirements of competition are Robustness, Fidelity or Invisibility, Capacity. The decoder or the detector consists of Blind as well as Non-blind. Some other factors involved are multiple watermarks, security, and cost. Evaluation of the system is also required. Multiple watermarks are needed for various purposes for different applications. The cost also depends on the application (Kaur et al., 2020). Speed and detectors deployed are the main issues taken into concern.
Security must have the ability for resisting hostile attacks that affect watermarks. Removal of unauthorized data is also required.
Security involves forgery as well as detection of unauthorized data.
Every security is dependent on keys. Cryptology assumptions state that adversaries have every information except about keys. It is important the hide the locations where watermarks are embedded (Chen et al.,2020).
2D signals consist of both vertical and horizontal dimensions. There are matrices in MATLAB. There are different formats like jpeg, gif, png, BMP and tiff, etc. video signals require 3D signals. Text, audio, and image all constitute multimedia. There are special formats of different files. Multimedia is represented in the computers as numbers. MATLAB modifies data and stores them with specific formats. It has a special toolbox and commands.
Two types of transforms are DFT and DCT. The computer needs a binary system to represent Audio signals or frame pixels. MSB is related to the content directly. LSB does not affect content, rather it represents noise, as it is redundant. Frequency components are summarised through DFT and DCT.
In a noisy channel, communication can be done better using watermarks. Techniques can be borrowed from communication theories that are established. Pixel domain involves blind embedding. The cover owner generates a Gaussian Noise that is known as Watermark. The original work adds this and the cover is blind. The decoder side denotes the linear relation between the work received and the watermark.
In some cases, the watermark cannot be recovered using blind embedding as the relation between cover and watermark is quite strong. So the image of the cover should be taken into consideration to solve this problem.
Haar wavelets are mostly simple and HVS is not reflected. DWT's earliest methods were based on WM.
B. LSB and DCT based watermarking
LSB stands for Least Significant Bit and it is the first bit on the right side.
Original reset LSB
If both the LSB and watermark bits are the same, there is no change needed. If both are different, the LSB bit is set the same as the watermark bit. A little distortion is required for the perfect visibility of the host image. This cruises violation of robustness. Different signals provide better representation. They are more concordant to HVS (Naseeret al., 2019).
DCT (discrete cosine transform) is similar to DFT. it is required to represent the domain of frequency. Only real numbers are used. This is mainly used for the compression of images as well as video. Two types of DCT are DCT in 2D and inverse DCT. Log functions are used to examine the coefficients as low frequencies have larger values. DCT is the worst in the case of spatial localization for any particular frequency (Fares et al., 2021).
A. Tsai et al proposed watermarking method
This method is based on fast DCT for implementing the processor using the digital signal. The host image is not always necessary for embedding and efficient extraction. Watermarking keys processes require 4 coefficients of frequency in DCT. To normalize the watermark, 2 permutation vectors are required which are taken randomly, and also a quantized matrix is required. To decrease 2D image complexity and to reduce the time to 0.33 seconds, a fast DCT algorithm is required. Excellent quality watermark images are extracted. This generally works on the process of watermarking that is reversible. It is based on histogram shifting that helps in the prediction of the errors. Several predictions are designed instead of one prediction error, so that multiple errors can be detected in a single time, using a new scheme. These errors refer to the asymmetric histograms of errors that are constructed by choosing a suitable error among all. The use of this histogram instead of the symmetric one decreases the number of pixels that are shifted. This helps in improving the quality of the image that is watermarked. Minimum as well maximum error histograms are combined and this strategy is used as a complementary one for embedding. Some of the pixels of the image that watermarked can be restored if the error histograms are shifted in opposite direction (Su et al., 2021). Experiments show these methods remake images that are of high quality and that are watermarked, they have a large embedding capacity (Abdulazeezet al., 2021).
B. Fu et al proposed watermarking method
This technique mainly involves the robustness of the watermarking that is needed for textured images. The growing internet with lots of multimedia applications gives a rise to various medical science as well as defense applications issues. The problems related to privacy can be solved by introducing watermarking methods. Robust watermarking is based on the matching of texture with the host image. Watermark is embedded in various places of the image to snake the security stronger. The NCC and PSNR algorithms are required in the testing of the efficiency of the watermarking. Geometric distortion is a way for watermarking images. This method is based on the feature of detection of images (Molina-Garciaet al., 2020). Local regions are needed for the embedding. During the end of the detection, the hopeful points are robustly detected as well as used for synchronization again. Facial features were one of the earliest methods for exploiting the previous idea on which it was based. The most important features for the image to be watermarked were eyes as well as mouth. Features of the face were localized and were used for the geometric charges. The watermarking method that is involved in the multiwavelet domain is based on SVMs. these methods mainly work on the frequency band and image properties. A mean value method for modulation is needed for performing the decomposition of multiwavelet in single-level on blocks of every image. The modulation methods are needed to reduce the distortion of any image. SVMs learn the relationship at the watermark detector. SVMs have very good learning abilities. So these are required to extract the watermarks correctly even if gone through various attacks.
DRM is used for financial transactions. The technologies that are needed for DRM are passwords, protection of network, encryption, fingerprinting, as well as detection of the copying process. REL is used for expressing language for digital content that is contractual. Digital contents are controlled by REL. ODRL provides a flexible model based on information, vocabulary, and the mechanisms of encoding that represent statements of usage of services. The hiding of information mainly is divided into two parts that are watermarking as well as steganography. Some international organizations that work on the watermarking process are CPTWG, SDMI, TALISMAN, etc. the advertisers ensure what they receive by purchasing from the broadcasters. Performers ensure the royalties they get from advertising businesses. Owners ensure their works are not pirated. The types of broadcast monitoring are the monitoring done by humans, monitoring passively, and monitoring actively. Fingerprinting is involved in every legal sale. Content authentication extracts signatures from a special work part. If a watermark is fragile it can be destroyed in any mode. The process of embedding involves effectiveness and fidelity as well as payload. The process of detection involves blind detection, robustness as well as false-positive behavior. The detector gives a positive response to the watermark. Effectiveness is the probability that states the output of the embedder is watermarked. These can be determined by empirical and analytical methods. Similarity can be identified through fidelity between the cover work as well as the counterpart of the watermark. Data payload is also called Capacity. It is dependent on work nature. FPR is expected to occur in multiple runs of the detector. FPR is dependent on the application, which is in hand. Robustness includes common image, video, and audio distortions. Watermarks cannot resist every type of distortion. The probable hostile attacks need special care. A series of waves carrying sounds, messages, and pictures is called a signal. It is of varying quantities. Audio is 1 dimensional. The computers work based on discrete signals. Digital sound’s can be stored in various formats like mp3 or Wav. Frequency sampling is also very important. Joseph Fourier introduced DFT. it helps in decomposing a signal into frequency. It gives various information about a particular signal. Any kind of signal can be expressed as the total sum of the sines and cosines by multiplying with the weighting function. It is applicable for even complicated signals also. Signals are held by MATLAB. The types of channels are fading channels as well as additive noise channels (Ahmadiet al., 2020). There are two types of detectors called blind detectors as well as the informed detectors. In a blind detector, there is no availability of the original image. There should be zero mean as well as a unit variance in reference watermark. Watermark can be newly interpreted by using communication. Embedder takes advantage of the availability of the original image. Cw can be set to any value using embedder. The images are converted to either FFT or DCT for embedding the watermark( Waqas et al.,2020).
Watermark is very much important for the security and protection of any kind of multimedia content. Due to the evolution and technology, there are many ways intrigued today for the protection of the various data contents. Embedding a watermark is one of them. There are many types of watermark methods that are introduced now for maintaining security. Some general attacks that are considered during the experiments for watermarking processes are scaling, dithering, JPEG coding, cropping, re-watermarking, print-scan as well as collusion. Some useful strategies of embedding are frequency or spatial techniques(Shehab et al.,2018). Watermark requires a Gaussian sequence. Cross-correlation is needed as a hierarchical search regarding the watermark.
Babu, R.G., Elangovan, K., Maurya, S. and Karthika, P., 2021. Multimedia Security and Privacy on Real-Time Behavioral Monitoring in Machine Learning IoT Application Using Big Data Analytics. In Multimedia Technologies in the Internet of Things Environment (pp. 137-156). Springer, Singapore.
Giri, K.J., Parah, S.A., Bashir, R. and Muhammad, K., Multimedia Security.
Bhat, M.I. and Giri, K.J., 2021. Impact of Computational Power on Cryptography. In Multimedia Security (pp. 45-88). Springer, Singapore.
Hosny, K.M. ed., 2020. Multimedia Security Using Chaotic Maps: Principles and Methodologies. Springer.
Kaur, B., Singh, D. and Roy, P.P., 2020. A study of EEG for enterprise multimedia security. Multimedia Tools and Applications, 79(15), pp.10805-10823.
Chen, J.I.Z. and Smys, S., 2020. Social Multimedia Security and Suspicious Activity Detection in SDN using Hybrid Deep Learning Technique. Journal of Information Technology, 2(02), pp.108-115.
Naseer, Y., Shah, D. and Shah, T., 2019. A novel approach to improve multimedia security utilizing 3D mixed chaotic map. Microprocessors and Microsystems, 65, pp.1-6.
Fares, K., Khaldi, A., Redouane, K. and Salah, E., 2021. DCT & DWT based watermarking scheme for medical information security. Biomedical Signal Processing and Control, 66, p.102403.
Su, G.D., Chang, C.C. and Chen, C.C., 2021. A hybrid-Sudoku based fragile watermarking scheme for image tampering detection. Multimedia Tools and Applications, pp.1-23.
Abdulazeez, A.M., Hajy, D.M., Zeebaree, D.Q. and Zebari, D.A., 2021. Robust watermarking scheme based LWT and SVD using artificial bee colony optimization. Indonesian Journal of Electrical Engineering and Computer Science, 21(2), pp.1218-1229.
Molina-Garcia, J., Garcia-Salgado, B.P., Ponomaryov, V., Reyes-Reyes, R., Sadovnychiy, S. and Cruz-Ramos, C., 2020. An effective fragile watermarking scheme for color image tampering detection and self-recovery. Signal Processing: Image Communication, 81, p.115725.
Ahmadi, S.B.B., Zhang, G., Wei, S. and Boukela, L., 2020. An intelligent and blind image watermarking scheme based on hybrid SVD transforms using human visual system characteristics. The Visual Computer, pp.1-25.
Waqas, U.A., Khan, M. and Batool, S.I., 2020. A new watermarking scheme based on Daubechies wavelet and chaotic map for quick response code images. Multimedia Tools and Applications, 79(9), pp.6891-6914.
Shehab, A., Elhoseny, M., Muhammad, K., Sangaiah, A.K., Yang, P., Huang, H. and Hou, G., 2018. Secure and robust fragile watermarking scheme for medical images. IEEE Access, 6, pp.10269-10278.