A Technique Steganography for Hiding Secret Information and its Application

Transcription

1 4 A Technique Steganography for Hiding Secret Information and its Application Jayshri Dnyaneshwar Pande 1, Dr. Trapti Arjeria 1 Bhabha Engineering Research Institute Bhopal,Computer Science Engineering, Bhabha Engineering Research Institute Bhopal, Computer Science Engineering ABSTRACT In this paper we have represented the technique steganography for hiding secret information. We have presented a high capacity and high stego-signal quality audio steganography scheme based on samples comparison in DWT domain where selected coefficient of a segment are compared with pre determined threshold value T and based on comparison bits are embedded. The strength of our algorithm is depend on the segment size and their strength are enabled the algorithm to achieve very high embedding capacity for different data type that can reach up to 5% from the input audio file size with lest of 35 db SR for the output stego signal. The proposed algorithm was implemented by using Matlab (009a) programming Keywords Steganography, audio, data hiding, coding, matlab. I. ITRODUCTIO Steganography is an art and a science of communicating in a way, which hides the existence of the communication. It is also called as covered writing, because it uses a cover of a message for sending any important secret message [3]. Steganography serves as a means for private, secure and sometimes malicious communication. Steganography is the art to hide the very presence of communication by embedding the secret message into the innocuous looking cover media objects, such as images using the human's visual, aural redundance or media objects' statistical redundance. Steganography is a powerful tool which increases security in data transferring and archiving. In the steganographic scenario, the secret data is first concealed within another object which is called cover object, to form stego object and then this new object can be transmitted or saved. Using different techniques, we can send secret data in the form of an image, a music file or even a video file by embedding it into the carrier, forming a stego signal. At the receiver s end, the secret data can be recovered from the stego signal using different algorithms [3]. II. AUDIO STEGAOGRAPHY In a computer-based audio steganography system, secret messages are embedded in digital sound. The secret message is embedded by slightly altering the binary sequence of a sound file. Existing audio steganography software can embed messages in WAV, AU, and even MP3 sound files. Embedding secret messages in digital sound is usually a more difficult process than embedding messages in other media, such as digital images. In order to conceal secret messages successfully, a variety of methods for embedding information in digital audio have been introduced [17]. LSB Encoding: Sampling technique followed by Quantization converts analog audio signal to digital binary sequence. In this technique LSB of binary sequence of each sample of digitized audio file is replaced with binary equivalent of secret message. Parity Coding: Instead of breaking a signal down into individual samples, the parity coding method breaks a signal down into separate regions of samples and encodes each bit from the secret message in a sample region s parity bit. Phase coding: Human Auditory System (HAS) can t recognize the phase change in audio signal as easy it can recognize noise in the signal. Spread spectrum: In the context of audio steganography, the basic spread spectrum (SS) method attempts to spread secret information across the audio signals frequency spectrum as much as possible. Echo Hiding: In echo hiding, information is embedded in a sound file by introducing an echo into the discrete signal. III. AUDIO STEGAOGRAPHY TECHIQUES A. An Overview: An audio steganography technique can be classified into two groups based on the domain of operation. One type is time domain technique and the other is transformation based method. The time domain techniques include methods where the embedding is performed without any transformation. Steganography is employed on the original samples of the audio signal.

2 5 One of the examples of time domain steganography technique is the least significant bit (LSB) method. In LSB method the watermark is embedded into the least significant bits of the host signal. As against these techniques, the transformation based steganography methods perform steganography in the transformation domain. Few transformation techniques that can be used are discrete cosine transform and discrete wavelet transform. In transformation based approaches the embedding is done on the samples of the host signal after they are transformed. Using of transformation based techniques provides additional information about the signal. In general, the time domain techniques provide least robustness as a simple low pass filtering can remove the watermark. Hence time domain techniques are not advisable for the applications such as copyright protection and airline traffic monitoring; however, it can be used in applications like proving ownership and medical applications [0]. B. LSB coding: Least significant bit (LSB) coding is the simplest way to embed information in a digital audio file. By substituting the least significant bit of each sampling point with a binary message, LSB coding allows for a large amount of data to be encoded. The following diagram illustrates how the message 'HEY' is encoded in a 16-bit CD quality sample using the LSB method [10]: Figure 1 Example of LSB coding C. Phase Coding: Phase coding addresses the disadvantages of the noise-inducing methods of audio Steganography. Phase coding relies on the fact that the phase components of sound are not as perceptible to the human ear as noise is. Rather than introducing perturbations, the technique encodes the message bits as phase shifts in the phase spectrum of a digital signal, achieving an inaudible encoding in terms of signal-to-perceived noise ratio. The phase coding method breaks down the sound file into a series of segments. A Discrete Fourier Transform (DFT) is applied to each segment to create a matrix of the phase and magnitude. The phase difference between each segment is calculated, the first segment (s0) has an artificial absolute phase of p0 created, and all other segments have newly created phase frames. Figure Phase shift coding The new phase and original magnitude are combined to get the new segment, Sn. These new segments are then concatenated to create the encoded output and the frequency remains preserved. In order to decode the hidden information the receiver must know the length of the segments and the data interval used. The first segment is detected as a 0 or a 1 and this indicates where the message starts. D. Echo Hiding: Echo hiding embeds its data by creating an echo to the source audio. Three parameters of this artificial echo are used to hide the embedded data, the delay, the decay rate and the initial amplitude. As the delay between the original source audio and the echo decrease it becomes harder for the human ear to distinguish between the two signals until eventually a created carrier sound s echo is just heard as extra resonance. In addition, offset is varied to represent the binary message to be encoded. One offset value represents a binary one, and a second offset value represents a binary zero. If only one echo was produced from the original signal, only one bit of information could be encoded. Therefore, the original signal is broken down into blocks before the encoding process begins. Once the encoding

3 6 process is completed, the blocks are concatenated back together to create the final signal. The embedding process is divided into the individual blocks such as encryption, Segmentation wavelet decomposition, frames selection, watermark embedding and reconstruction as shown in Figure 4 Audio Cover Signal Segmentation Wavelet Decomposition Coefficient Selection Embedding Reconstruction Audio Stego Signal Figure 3 Example of Echo hiding The "one" echo signal is then multiplied by the "one" mixer signal and the "zero" echo signal is multiplied by the "zero" mixer signal. Then the two results are added together to get the final signal. The final signal is less abrupt than the one obtained using the first echo hiding implementation. This is because the two mixer echoes are complements of each other and that ramp transitions are used within each signal. These two characteristics of the mixer signals produce smoother transitions between echoes. IV. PROPOSED METHODOLOGY In the proposed method the carrier file is taken as audio format and the secret message may be a text or audio format files. Our system provides a very friendly User Interface where the user had to specify just the required inputs (audio, text).after embedding or extracting the user can save /open or just discord the output of that particular operation according to their wish. In view of providing security by preventing unauthorized person to access the software password facility is provided to the user in order to work with the software. To provide more security by avoiding an intruder to extract the embedded data a security key is used while embedding and extracting message. There are two methods in Audio steganography 1) Embedding ) Extracting Embedding is a process of hiding the message in the audio. Extracting is a process of retrieving the message from the audio. EMBEDDIG ALGORITHM Secret Encryption Figure 4 the General Structure of the Proposed hiding Scheme Input secret message and cover signal: Proposed method starts by inputting the secret message which is to be embedding into signal. The secret message can be any text file or image or any audio wave file.and then inputting the cover signal in which data is to be embedded. This cover signal must be sufficient large to cover the message. After selection of input secret message and cover signal next, we find out the length of the audio file as well as length of the text file. Check whether the size of the audio file is greater or less than the text file. If the size of the audio file is less than the size of the selected text file then print the error message, otherwise it is possible to embed the text file into selected audio file. Encryption: Before hiding the secret message into cover signal it must be converted into the other form so that it can t be interpretable by intruder.to do so first, we convert the secret data or message into it s binary form.let suppose the length of message is bits long, ext use the random number to generate the private key of length same as the length of message because the size of encrypt message is equal to the original message, then apply X-OR operator to generate the cipher message of length bits. Cover Signal Segmentation: Let the input cover signal consist of R samples, this signal is segmented into two catagories:

4 7 1. Processed samples. Unprocessed samples. The size of processed samples is depending on the size of message bits. If the size of message bit is then The Processed samples are consist of *^L samples. Where L is decomposition level. the rest samples is called unprocessed samples, ext the processed part is partitioned into segments of size same as size of message bits that is segments; each segment has length of Z samples. EXRACTIG ALGORITHM The extraction process is illustrated in Figure 5. The extraction process is divided into blocks stego segmentation, wave decomposition, Coefficient selection, extraction and reverse encryption. The Quantization parameter Q needs to be the same that is used during encryption. Audio Stego Signal Segmentation Wavelet Decomposition Coefficient Selection Extraction Reverse Encryption Secret Figure 5 Block diagram of the Recovery Algorithm Input Audio stego signal: In the message recovery algorithm, first we select the Audio stego signal from which data is to be extracted. Stegno Signal Segmentation: Again, the stego signal is segmented into two categories: 1. Processed samples. Unprocessed samples. The size of Processed segment is known to receiver with the help of size of message bit.it is calculate by multiplying the size of message bits with ^L where L is the decomposition level. ext the Processed part is segmented again into segments; each segment has length of Z samples. Stego Segment Decomposition and coefficient selection: Again, each segment of the Stego audio signal is decomposed using L level of Haar DWT to obtain ^L signals, each one of the produced signal has length of Z/^L samples. One represents the Approximated signal and the others represent detailed signals. From the detailed signal we select same detail signal that was selected on the time of hiding stage for embedding process Secret Recovery Stage: Secret message recovery stage is very simple and based on comparison of selected detailed coefficient with threshold value T. If the coefficient is greater then or equal to threshold T it means that bit is 1 otherwise the bit is 0 if(segment(i,p)>=t) (=1; elseif(segment(i,p)<t) (=0; else error("there is problem in Stegno Signal"); end Reverse Encryption: Before delivery of the secret message to receiver, it must be converted back to it s original form.to do so first, we use the same random number to generate the private key of length same as the length of message, then apply X-OR operator to get the original message of length bits. ext we convert message of it s binary form to text format.and then deliver to the receiver. V. RESULT AD AALYSIS Performance Parameter: In this section we give brief descriptions of the performance parameters used. The performance parameters used for the performance are bit error rate (BER) and signal to noise ratio (SR) discussed below. The original signal (the cover document) is denoted x (, i 1,.. while the distorted signal (the stego-document) as y (, i 1,... In some cases the distortion is calculated from the overall data. However most of the case, the distortion is calculated for small segments and by averaging these, the overall measure is obtained. Bit Error Rate: Bit error rate can be defined as the percentage of bits corrupted in the transmission of digital information due to the effects of noise, interference and distortion. For example, the bits to be transmitted are and the received bits are Comparing the number of bits transmitted to received, two bits are affected by transmission. Hence, the BER in this example is /8*100 = 5%. Generally the BER of a signal is computed using equation below. BER= W / S *100 Where, w is the number of error bits and S refers to the o of Secret bits

5 8 Signal-to-oise Ratio: SRseg is defined as the average of the SR values over short segments: 10 M SRseg M 1 m0 log 10 m 1 im x ( x( y( where x( is the original audio signal, y( is the distorted audio signal. The length of segments is typically 15 to 0 ms for speech. The SRseg is applied for frames which have energy above a specified threshold in order to avoid silence regions. Signal-to- oise Ratio (SR), is a special case of SRseg, when M=1 and one segment encompasses the whole record [18]. The SR is very sensitive to the time alignment of the original and distorted audio signal. The SR is measured as SR 10 log 10 x( y( i1 i1 x ( Here represents the number of samples in both signals. Experiment Setup: All algorithms, including proposed technique, are implemented on Windows PC having Intel.4 GHz processor and GB RAM, and run using Matlab 9a. We have considered three different audio files in this experiment to embed digital data. One of the audio file is adios sound and is a 8 bit mono audio signal sampled at 11.5 khz. We applied Haar wavelets on cover signal and choose the coefficient where the data is to be hide using a pre determined threshold value T. The performance of the embedded information is studied by applying attacks such as re-quantization, re-sampling, low-pass filtering, high-pass filtering, AWG, MP3 compression, jittering and cropping. For the complete analysis of the proposed technique different audio signals are considered such as the adios, aaaaagh, and shutdwn track namely, and respectively. Figure 6 shows the time domain response of these signals. Care has been taken to study the complete performance of the algorithm by collecting diverge audio signals as shown in Figure 6. Same attacks are employed on all audio signals. Figure 6 Time domain response of the considered audio signals Performance Analysis: Fig.7 shows the relationship between SR and embedding capacity for fixed message type and three different cover signals. SR(db) SR V/S Data Rate Data Rate(KBPS) Figure 7 The Relationship between SR and Embedding Capacity for Different Cover Signals and different Data Type The performance of the proposed algorithm against the signal processing and desynchronized attacks such as the addition of Gaussian noise is evaluated. The wavelet filter used for the analysis is Hear wavelet with level 3 decomposition and the SR of the stego signal is 34 db and T= The data taken is a text file with 3kb size.

6 9 From the observation, for SR value of 5 db and above the embedded data is inaudible to the human ear. In addition to that, the BER of the extracted data is nearly 1 and 0 for majority of the attacks. Table 1 Performance evaluation of the embedded data with respect to AWG and BER Cover Signal AWG (DB) BER Table 1 shows a comparison of different cover signal with respect to SR on different values of Threshold T and processing time for fixed capacity ( about 00 word/sec) and Z = 8 samples. In these tests we use adios, aaaaagh, and shutdwn wave file as a cover signals with length of samples and respectively and text file as a secret message with size of 3kb. The results in table shows that using the threshold value T equal to 0.1 for comparison will increase the SR. The arbitrary result of bits block matching make the distribution of secret message blocks over the cover signals arbitrary and that increase the security of secret message. Fig 8 shows a comparison graph of different cover signal with respect to SR on different values of T and processing time for fixed capacity ( about 00 word/sec) and Z = 8 samples. The comparison showed the clearly superiority of the proposed scheme over the conventional DWT scheme in high embedded capacity, the SR is above 5 db in our algorithm while it is in range of 1 db in conventional DWT scheme for different data type messages. db SR in (db) Figure 8 Comparison Value graph of P for and different Q cover signals with respect to SR on different Threshold T Fig 9 shows a comparison graph of different cover signal with respect to Processing Time on different values of T for fixed capacity ( about 00 word/sec) and Z = 8 samples. The comparison showed the clearly superiority of the proposed scheme over the conventional DWT scheme in high embedded capacity, the SR is above 5 db in our algorithm while it is in range of 1 db in conventional DWT scheme for different data type message Second Processing Time (4,5) (3,6) (,7) Value of P and Q Figure 9 Comparison graph for different cover signals with respect to Processing Time on different value of T VI. COCLUSIO Steganography is an information hiding technique where secret message is embedded into unsuspicious cover signal. An effective audio steganographic scheme should possess the following three characteristics: Inaudibility of distortion (Perceptual Transparency), Data Rate (Capacity) and Robustness. These characteristics (requirements) are called the magic triangle for data hiding.

7 10 We have presented a high capacity and high stego-signal quality audio steganography scheme based on samples comparison in DWT domain where selected coefficient of a segment are compared with pre determined threshold value T and based on comparison bits are embedded. The strength of our algorithm is depend on the segment size and their strength are enabled the algorithm to achieve very high embedding capacity for different data type that can reach up to 5% from the input audio file size with lest of 35 db SR for the output stego signal. The proposed algorithm was implemented by using Matlab (009a) programming. The proposed algorithm was tested using three audio cover signals: adios, aaaaagh, and shutdwn wave file called, and respectively. Each signal has resolution of 8 bits per sample and sampling frequency 1105 samples/sec and text are used in tests as secret messages. The quality of output signal in each test was computed using SR. Disadvantages associated with this proposed system are a low data transmission rate due to the fact that the each bit of secret message is embedded into one segment and size of segment is approx 4 to 16 samples. It means that utilization of samples is very poor. As a result, this method can be used when only a small amount of data needs to be concealed. Otherwise this can be proved as a good method for audio Steganography. In future we will modify and improve this technique so that more data can be embedded into cover signal. REFERECES [1] Zaidoon Kh. AL-Ani, A.A.Zaidan, B.B.Zaidan and Hamdan. O. Alanazi, Overview: Main Fundamentals for Steganography, journal of computing, volume, issue 3, march 010, issn [] Sos S. Agaian, David Akopian and Sunil A. D Souza TWO ALGORITHMS I DIGITAL AUDIO STEGAOGRAPHY USIG QUATIZED FREQUECY DOMAI EMBEDDIG AD REVERSIBLE ITEGER TRASFORMS 1on-linear Signal Processing Lab, University of Texas at San Antonio, Texas 7849, USA. [3] Ashwini Mane, Gajanan Galshetwar, Amutha Jeyakumar, DATA HIDIG TECHIQUE: AUDIO STEGAOGRAPHYUSIG LSB TECHIQUE International Journal of Engineering Research and Applications (IJERA), Vol., Issue 3, May-Jun 01, pp [4] eil Jenkins, Jean Everson Martina, Steganography in Audio Anais do IX Simpósio Brasileiro em Segurança da Informação e de Sistemas Computacionais page: 69-78,007 [5] Westfeld, A. (003). Detecting low embedding rates. In Petitcolas, F. A., editor, Information Hiding: 5th International Workshop. Springer. [6] I. J. Cox, J. Kilian, F. T. Leighton and T. Shamoon, Secure Spread Spectrum Watermarking for Multimedia, IEEE Trans. Signal Processing, vol. 6, no. 1, pp , December [7] B. Chen and G. W. Wornell, Digital Watermarking and Information Embedding using Dither Modulation, Multimedia Signal Processing, 1998 IEEE Second Workshop, pp: 73-78, December [8] edeljko Cvejic, Algorithms For Audio Watermarking And Steganography, Department of Electrical and Information Engineering, Information Processing Laboratory, University of Oulu 004. [9] Petitcolas, F.A.P., Anderson, R., Kuhn, M.G.,"Information Hiding - A Survey", July1999. [10] Prof. Samir Kumar Bandyopadhyay, Tuhin Utsab Paul, Avishek Raychoudhury, A Robust Audio Steganographic Technique based on Phase Shifting and Psycho acoustic Persistence of Human Hearing Ability, International Journal of Computing and Corporate Research. [11] Inas Jawad Kadhim, A ew Audio Steganography System Based on Auto-Key Generator, Al Khwarizmi Engineering Journal, Vol. 8, o. 1, PP 7-36 (01). [1] Mazdak, Z., A.M. Azizah, B.A. Rabiah, M.Z. Akram and A., Shahidan A Secure audio steganography approach, World Acad. Sci. Eng., Technol., 5: , 009. [13] Pal S.K., Saxena P. K. and Mutto S.K. The Future of Audio Steganography. Pacific Rim Workshop on Digital Steganography, Japan, 00. [14] Lee, Y. K. and Chen L. H. High Capacity Image Steganographic Model. IEEE Proceedings Vision, Image and Signal Processing, pp , 000. [15] Steganography, d14_gci ,00. html [16] Ronak Doshi, Pratik Jain, Lalit Gupta, Steganography and Its Applications in Security, International Journal of Modern Engineering Research (IJMER), Vol., Issue.6, ov-dec. 01 pp [17] International Journal of Electronics Communication and Computer Technology, Audio Steganography in a utshell, Issue 5 (September 01) ISS: [18] Abdulaleem Z. Al-Othmani1, Azizah Abdul Manaf and Akram M. Zeki3, A Survey on Steganography Techniques in Real Time Audio Signals and Evaluation IJCSI International Journal of Computer Science Issues, Vol. 9, Issue 1, o 1, January 01. [19] Johnson, eil F., Steganography, [0] edeljko Cvejic, Tapio Seppben Increasing the capacity of LSB-based audio steganography FI University of Oulu, Finland,00.