Abstract: In this study, projected a model for safe secret-digital file transmission by having Steganography as the ground. Steganography, in prehistoric times, dealt with camouflage of secrets between two parties. A part from text, many images, audio and video in millions are shared between millions. All sorts of the aforesaid face a common and serious challenge. Here proposed three routines, in which clandestine information is embedded based on the key in method 1, i.e., number and position of 1s in the four Least Significant Bits (LSB) of the key decides the pixels for embedding. In method 2, position of 1s in the entire key decides the pixels and the position of 1s in four LSBs only decides the location for embedding. In method 3, secret data is plunged by means of the same process but not directly. Instead, it is encoded by Huffman Coding and then embedded. Justification for this study is given by computing Mean Squared Error (MSE) and Peak Signal-to-Noise Ratio (PSNR) values for all the planes in the images by taking sample key values for all the three methods. Experimental results are demonstrated and the corresponding image outputs are also presented.
INTRODUCTION
From the famous quote “actions speak louder than words ”. Present days communication deals with the means to communicate the human actions directly to others rather than mere words or voices. Technology has been carved into many shapes over the years and will still have potential of a dramatic increase. Communication by actions involves the digital media (Cheddad et al., 2010), which is reforming and revolutionizing its dynamism every moment. This growth of the technology in the communication field have plethora of threats that are going to diminish the growth (Stefan and Fabin, 2000). With the birth of technology growth, took birth the hackers who easily hunt for the data and gather by some means. This scared many of the inventors for getting better technology into implementation (Amirtharajan and Rayappan, 2012a, b; Hmood et al., 2010a, b; Thenmozhi et al., 2012).
Security concerns and threats posed by the hackers are being tackled by algorithms which are capable of completely distorting the information making it practically undiscoverable for any third person. This means of distortion of information is coined cryptography (Schneier, 2007). Encrypting algorithms are the base for this method. They operate on the piece of data and change its complete form and outlook from its root level (the bits).
But Cryptography ’s encryption policy being too transparent lead many complications (Zaidan et al., 2010), as once anyone sees anything distorted; it ’s evident that something is being hidden. Hence arose, steganography with maximum opaqueness in its hiding methods. A cover that is used in this method has the capability of practically hiding anything like image, video, sound etc (Amirtharajan et al., 2012; Al-Frajat et al., 2010; Zhu et al., 2011) and specialty being the protection of added randomizations from the intruders apart from the source. This has made steganography very popular among masses and is attracting many towards it for further developments. Though having such great security tool, one must not be cautious enough to correctly use it, nor might it end up similar to the September ’11 and destruct the mankind for such a tool. Hence this study suggests three methods for random image steganography to improve imperceptibility.
PROPOSED METHODOLOGY
General schematic diagram of the proposed is given in Fig. 1.
Method 1: This method embeds secret data into the pixels as per the key ’s four LSBs.
If key K = 13[0 0 0 0 1 1 0 1] then embedding should be in the 1st (2^0), 3rd (2^2) and 4th (2^3) LSBs. If key = 5[0 0 0 0 0 1 0 1] then embedding should be in the 1st (2^0), 3rd(2^2) LSBs.
Method 2: It will embed data into the selected pixels based on the key. And also this method will embed data into the selected pixels based on the four LSBs of the Key.
If key K = 173 [1 0 1 0 1 1 0 1] then embedding should be done in the pixels1, 3, 5, 6, 8. This sequence of embedding should be repeated for a block of eight pixels for complete embedding. In a each selected pixel the embedding should be in the 1st (2^0), 3rd (2^2) and 4th (2^3) LSBs. If key K = 165 [1 0 1 0 0 1 0 1] then embedding should be done in the pixels1, 3, 6, 8. This sequence of embedding should be repeated for a block of eight pixels for complete embedding. In each selected pixel the embedding should be in the 1st (2^0), 3rd (2^2) LSBs.
| Fig. 1: | Block Diagram of proposed system |
| Fig. 2: | Flow chart for method 1 |
Method 3: First data ’s are embedded using Huffman code then embed encoded data into the selected pixels based on the key. And also this method will embed encoded data into the selected pixels based on the four LSBs of the Key. If key K = 173 [1 0 1 0 1 1 0 1] then embedding should be done in the pixels, 3, 5, 6, 8. This sequence of embedding should be repeated for a block of eight pixels for complete embedding. In a each selected pixel the embedding should be in the 1st (2^0), 3rd(2^2) and 4th(2^3)LSBs. If key K = 165 [1 0 1 0 0 1 0 1] then embedding should be done in the pixels 1, 3, 6, 8. This sequence of embedding should be repeated for a block of eight pixels for complete embedding. In each selected pixel the embedding should be in the 1st (2^0), 3rd (2^2) LSBs.
Algorithms for all three methods are explained below and their corresponding flowcharts are given in Fig. 2-4.
Algorithm for method 1:| Fig. 3: | Flow chart for method 2 |
| Fig. 4: | Flow chart for method 3 |
RESULTS AND DISCUSSION
In this study, color images Lena and rose 256x256x3 are taken as covers for the three methods as shown in the Fig. 5a, 6a, 7a, 8a, 9a,10a. For key value of 205, the stego images are given in Fig. 5b, 6b, 7b, 8b, 9b, 10b. The MSE and PSNR values of original and stego images are given in tables. There exist surplus methods for information hiding which aims for robustness, secrecy and imperceptibility. This article swathes all the specifications and boasts that it is steadfast and unswerving when compared with the others.
Method 1 key value = 205 (Fig. 5, 6), Method 2 key value = 205 (Fig. 7, 8) and Method 3 key value = 205 (Fig. 9, 10).
Image quality metrics
Mean squared error (MSE): The average squared difference between
a original image and resultant (stego) image is called Mean Squared Error. It
dampens small variation between the two pixels but reprimands large ones.
| Fig. 5(a-b): | Lena (a) Cover image and (b) Stego image |
| Fig. 6(a-b): | Rose (a) Cover image and (b) Stego image |
| Fig. 7(a-b): | Lena (a) Cover image and (b) Stego image |
| Fig. 8(a-b): | Rose (a) Cover image and (b) Stego image |
| Fig. 9(a-b): | Lena (a) Cover image and (b) Stego image |
| Fig. 10(a-b): | Rose (a) Cover image and (b) Stego image |
where, M and N denote the total number of pixels in the horizontal and the vertical dimensions of the image Ci,j represents the pixels in the cover image and si,j represents the pixels of the stego-image.
Peak signal-to-noise ratio (PSNR): The higher the PSNR, the nearer the stego image is to the cover. A higher PSNR value associates to a high quality image:
Table 1 and 2 are the corresponding MSE, PSNR values for method 1
Table 3 and 4 are the corresponding MSE, PSNR values for method 2.
Method 3: First data ’s are embedded using Huffman code then embed encoded data into the selected pixels based on the key. And also this method will embed encoded data into the selected pixels based on the four LSBs of the Key.
If key K = 173 [1 0 1 0 1 1 0 1] then embedding should be done in the pixels 1, 3, 5, 6, 8. This sequence of embedding should be repeated for a block of eight pixels for complete embedding. In a each selected pixel the embedding should be in the 1st (2^0), 3rd (2^2) and 4th (2^3) LSBs.
Table 5 and 6 gives the corresponding MSE, PSNR values for method 3.
All the stego images give better PSNR (Table 1-6) values for the methodologies and symmetric key plays a major role in retrieval of confidential information.
| Table 1: | MSE, PSNR values for method 1 for Lena image |
| Table 2: | MSE, PSNR values for method 1 for rose image |
| Table 3: | MSE, PSNR values for method 2 for Lena image |
| Table 4: | MSE, PSNR values for method 2 for rose image |
| Table 5: | MSE, PSNR values for method 3 for Lena image |
| Table 6: | MSE, PSNR values for method 3 for rose image |
CONCLUSION
Steganography is the gloomy cousin of cryptography, the use of codes. While cryptography fixes up privacy, steganography is intended to provide secrecy. The focal conventions of a steganographic plot are that it should offer high payload and should be complex enough for the interloper. Selection of pixels and LSB over cover image in order to hide data based on key plays an indispensable role. At the same time, visual artifacts should be nil; this is characterized by PSNR. This study satisfies all these criteria to a greater extent and offers anticipated results and it also transfigure version of other methods. Cover image used in this study helps to improve randomization during the time of embedding secret data; thus it is a type of security all the way through anonymity in view of the fact that no surreptitious data can be rejuvenated without the consent of the keys and coding tactics. The main aim of this study is providing security; security cannot exist in nature, so this study bestows security for confidential data by embedding it in a cover image with the help of this new embedding method. To conclude, this method has philanthropic commercial prospective.