Image security using digital watermarking

The primary objectives of entrenching this secret information to images are to develop robustness requirements, imperceptibility, intangibility and vigor requirements. The most commonly applied techniques and algorithms in image watermarking this essay will embark on are the Discrete Wavelet Transform (DWT) and Singular Value Decomposition (SVD) (Rahman, 2013) Introduction The evolution of internet and world web has played very critical roles in documents transmission. The advantages of this technology-based software have played are greatly appreciated in the recent world; in fact, it has transformed the world to be like a global village (Rahman, 2013). The documents transmitted may be inform of objects, data, videos and most importantly the images. Due to emergencies and the need to cut down the operational costs, many organizations are turning to digital data transmission among other reasons.

Digital watermarking simply refers to embedding invisible, secret, and secured information inside a given image. This act differentiates it from duplicate or pirated image hence a sense of originality and authenticity. Other benefits includes proof of ownership, it gives sense of integrity, fragility, copyright ownership of data and unambiguous (Singe et al, 2014) In the case of image security or watermark, the algorithms are grouped into one of the given two domains. These are the transform or the spatial domain. In the domain of spatial, the security data is inserted directly by adjusting the pixel measures of the intended image, while on transform domain, the security data is embedded by adjusting transform coefficients domain. This is very vital to the company since it differentiates its products from other companies in the same field.

For instance, a company may be engaging in production of digital images closely similar to the competitors, in such a scenario, the rival company may produce similar counterfeit digital images which appear similar when viewed with naked eyes. This poses a big challenge to the brand owner and thus the need to imbed a watermark to create a distinction. It is recommended to imbed a unique digital watermark that cannot easily be manipulated by hackers. Literature Review Various scholars have conducted different studies on watermarking process. The height of mw ought to be shorter than that of mc 3. The loop at the cover image will be; Mc=Mpixel [i, j] – Mpixel [i, j]%2 Mc=Mc+Mw%2 Mw=Mw%2 When we reach at the height and the width of the watermarked image, then the end will have been reached.

Least Significant Bit watermarking technique has both advantages and disadvantages. The advantageous part of it is that it is simple and the watermark is not visible through the naked eyes. The disadvantage of it is extremely sensitive and that any attack on the watermarked image may lead to destruction of the watermark hence damaging the authentication (Thanki et al, 2017). In hybrid watermarking, even without accessing information about the cover image, authentication is possible. Hybrid watermarking also involves the combination of Discrete Wavelet Transform (DWT) and the Singular Value Decomposition (SVD) (Thanki at el, 2017) Discrete Wavelet Transform Discrete Wavelet Transform (DWT) is applied in processing of signals applications, these are video and audio compression, sieving noise in an audio, and simulating wireless aerial distribution (Thanki et al, 2017) Wavelet transform utilizes wavelet filters in transforming the image.

There are various filters but the most common applied are the Daubechies Bi-Orthogonal filters and the Haar Wavelet Filter. These mentioned filters decompose the digital image into various frequencies (Rahman, 2013). Discrete Wavelet Transform decomposes the digital image using either of the above filters to decompose the digital image into rate channel of uniform bandwidth. (ii) the singular value relates to the fundamental algebraic image characteristics. The mathematical algorithm Embedding the watermark in the digital image The following are the steps to be followed in embedding a watermark by use of Haar wavelet filter 1. Note the size of the cover of digital image and the watermark 2. Apply the Haar Discrete Wavelet Transform technique to decompose the digital image to decompose the digital image into four equal sub-bands, that is LL, LH, HL, and HH 3.

Apply the Singular Value Decomposition (SVD) to the first sub-band (LH) as follows: IO=UOSOVOT Where IO represents the LH Thus, LH= UOSOVOT 4. The extracted embedded information is retrieved Imperceptibility of the digital image This refers to the apparent quality of host digital image that should not be inaccurate by the presence of watermark. The Peak Signal-to-Noise Ratio (PSNR) is commonly used for evaluating the watermarked digital image (Li et al 2017) It is illustrated in the formulae below PSNR = 10 log10 (255/MSE)2 dB MSE=∑i= 0 ∑j=0 [( aij-bi j )2 / n x m] Where m, x, and n is image size aij and bij are the corresponding pixel measures of the two digital images. To demonstrate the above approach, the image below will be regarded as an original and the copyright digital image will be used as the watermark.

Figure 1. 1 depicts the cover original digital image, figure 1. This is because it ensures that the information is hidden and a secrete message is concealed in digital domain. To improve the security part of the watermarking system is a challenge since to ensure a robust and imperceptible watermark is not so possible. The hidden information must be secure for effective security of the image or the video. A comparison between the Discrete Wavelet Transform (DWT) and Singular Value Decomposition (SDV) is done and its found that DWT is more advantageous than SDV since it contains both frequency domain and spatial domain. In addition the watermark is invisible hence insecure, however, the advantage of SDV over DWT is that a significance addition of agitation to the watermark leads to a more variation in its values.