A Novel Approach for Data Hiding Technique Using Interpolation And Braille Method

— Steganography means concealed writing of secrete information within any cover media. In image Steganography for the purpose of privacy of secrete message, it is embedded within an innocuous image without knowing the existence of information to the intruders. To maintain the privacy and in direction to maximize it, this proposed technique used the Braille method. Braille method is the interpreting technique to scan the words for the blind peoples. So, with regard to make the secrete information invisible to the intruder Braille method helps a lot. In the proposed technique message characters are converted into 6 bits Braille’s binary representation and then embed the message bits within the interpolated pixels of cover image. Interpolated cover image is hatched through the actual image by using interpolation method and all message bits are inserted within the interpolated pixels from left to right using the even odd pixel value approach with parity checker. The proposed technique maximizes the message hiding capacity and also improves the PSNR and MSE value as compared to existing techniques.

T. Bedwal and M. Kumar proposed more secured image Steganographic technique using RGB-box mapping that improves previous LSB method. In this technique, a RGB image has to be hidden within another RGB image based upon LSB insertion method. Mapping was applied on RGB image that provides more security to secrete data and also no need of knowledge of mapping techniques [7].
Rajkumar, Rahul Rishi and Sudhir Batra proposed a technique for gray level images in which Parity Checker used for insertion of message bits at LSB bits of pixels. In this technique, pseudo-random generator was used for insertion. To insert '0' as message bit, pixel bits must have odd Parity and to insert '1' as message bit, pixel bits must have even Parity. This method provided better results for message insertion in 1st chances are 98.82% with very less changes in original image [8].
Ki-Hyun et al., proposed semi-reversible Steganographic approach which was based upon interpolation and LSB insertion technique. Interpolation technique used to scale up original image and LSB technique used for insertion of secrete data that provided better results with larger capacity to hide and higher quality of image [9]. This method provides higher insertion capacity and invisible distortion in image.
K. Joshi and Rajkumar introduced an approach for hiding the data using XOR for image Steganography by using last three bits of pixels. Here XOR operation is performed on message bits and original image [10]. This method provides higher security to the data from intruders to detect.
Al-Hussien Seddik Saad et al., proposed a method using Zero-Order-Holding Method (ZOH) which resulted as higher PSNR value rather than other LSB methods. In this method, cover image resulted as zoomed out image using ZOH method and using LSB substitution, message is inserted. This technique provided a higher security as well as also maximizes message holding capacity (MHC) comparatively [12].
Al-Hussien Seddik Saad et al., proposed an improved technique for previous method as Zero order Holding, by combining LSBraille image Steganographic technique. In this method, firstly secrete message is converted in 6 bits binary representation using Braille method and then used LSB substitution and ZOH method to insert the message. ZOH Braille technique provided much higher PSNR values and increased MHC compared to the previous technique [13].
III. PROPOSED WORK In the proposed method, firstly modification is done in the confidential message by using Braille method which increase maximum hiding capacity rate. In the Braille (blind people language) method characters representation, each character is represented by 6 dots matrix known as Braille cells. Binary representation of these characters includes only 6 bits rather than 8 bits in ASCII binary representation which directly saves 2 pixels from every byte insertion procedure [15]. Braille Cell represents the characters for example character 'a' has dot 1 as dark only.

Braille Cell
Braille representation Of character 'a' These six dots consist of total 64 possible combinations. 6 bits representation of characters using Braille method is shown in table below where black dots are represented by 1 and empty dots are used by 0. A confidential message M is inserted within a cover image (CI). Cover image is generated from the original image using interpolation method which is discussed below.
 Take two adjacent pixels from original image (I), subtract minimum pixel value from maximum pixel value among both. P= MAX-MIN.  Divide P by 2 and take floor value (P=P/2).
 Add this value to the min pixel value. P=MIN+P.
 Then interpolated value is generated. If this procedure continues firstly row wise in matrix among each adjacent pixel then follow it column wise.  Now, resulted pixel matrix as the cover image (CI).
 Let M*N Pixels is the size of the cover image and after the interpolation cover image size becomes (2M-1)*(2N-1) Pixels. These interpolated pixels in cover image are used to hide the secrete message. Insertion and Extraction algorithms are further discussed here as below.

A. Insertion Algorithm
Step 1. Initialize original input image (I) of M*N Pixels. Generate cover image (CI) using above method and M is the confidential message which is in converted form of 6 bits by using Braille representation (m 1 , m 2 ….. m n ).
Step 2. Message bits are inserted in cover image interpolated pixels firstly row wise then column wise. For insertion of message take each two adjacent pixels from LSB of message M.
Step 3. If insertion message bits are '00'. Then check Pixel (P) value and its parity. If pixel value is even number with odd parity then no any change required in pixel and insert the message bit '00'. Else go to Step 7 and insert the message bits '00'.
Step 4. If insertion message bits are '10'. Then check Pixel (P) value and its parity. If pixel value is even number with even parity then no any change required in pixel and insert the message bit '10'. Else go to Step 7 and insert the message bits '10'.
Step 5. If insertion message bits are '11'. Then check Pixel (P) value and its parity. If pixel value is odd number with odd parity then no any change required in pixel and insert the message bit '11'. Else go to Step 7 and insert the message bits '11'.
Step 6. If insertion message bits are '01'. Then check Pixel (P) value and its parity. If pixel value is odd number with even parity then no any change required in pixel and insert the message bit '01'. Else go to Step 7 and insert the message bits '01'.
Step 8. Choose one from step 3 to step 6. Do these steps repeatedly until the length of message is covered or inserted in interpolated pixels.
Step 2. Extract the interpolated pixels.
Step 3. If Pixel value (P) is even number with odd parity bits then message bits are '00'.
Step 4. If Pixel value (P) is even number with even parity bits then message bits are '10'.
Step 5. If Pixel value (P) is odd number with odd parity bits then message bits are '11'.
Step 6. If Pixel value (P) is odd number with even parity bits then message bits are '01'.
Step 7. Choose from step 3 to step 6 according to requirement. Collect the message bits from these pixels from left to right. Do these steps repeatedly until the complete message bits extracted. After getting the complete extracted message bits convert these message bits into original message by using Braille's character representation table.
It can be understood through the example. Let us take an original image with size 4*4 pixels then generate the interpolated pixels by following above discussed procedure. We have a message M as 1011010100101101 and interpolated pixels generated as from left to right. This is the explanation part of this proposed algorithm which is discussed above with the help of an example. This algorithm is implemented using MATLAB on distinct datasets of images to analyse this proposed work.
IV. PARAMETERS USED FOR ANALYSIS PSNR: It stands for 'Peak Signal to Noise Ratio'. It is defined as the ratio of maximum possible power to the corrupting noise which affected the image representation. Unit to measure this factor is decibel (DB).

MSE:
It stands for 'Mean Square Error' which can be evaluated as the difference measuring function between original certified image pixels and stego image pixels. MSE is a risk function which can be calculated as expected squared error value. EXPERIMENTAL RESULT This proposed algorithm is evaluated to take its advantage over the already existing works. Different sample of images are tested for the comparison purpose of this algorithm to the existing algorithms.

VI. CONCLUSION
In image Steganography, different techniques have been already in uses which are proposed earlier by many of the researchers or authors with respect to the concealed message transformation between sender and receiver by obscuring the message within cover media i.e. image. In this perspective to enhance the security to the confidential information, this paper introduced a novel technique for obscuring the data. This paper provides a novel Steganography technique which uses Braille's method for message conversion and even odd pixels values and parity checker for message hiding purpose. It concludes with the higher capacity, PSNR and MSE values and future work may improve this technique.