Bio-hiding for Smart Swipe Card: A Secret Security

INTRODUCTION

Authentication using biometric involves obtaining physical traits from an individual like finger marks, iris, facial, voice and retinal models, written monikers etc., for certifying an individual’s identity via processing images. Important here is mentioning that characteristics obtained are unique to an individual and remain so in his life time. Finger-print based biometric authentication system has in this regard evolved as one of the oldest and most mature system due to the unique and reliable finger-print of a person and successfully implemented in many varied applications (Cheddad et al., 2008). The fingerprints analysis in favour of harmonizing intention usually involves likening of quite a few traits in the feature prototype. They comprise models, having amassed features like crumples (raised skin), furrows (lowered skin) and minutia points, which are distinctive traits that create inside such moulds. Indeed it becomes crucial of having knowledge on construction plus assets belonging to individual with the intention of effectively utilizing a number of technologies in imaging. First and foremost Minutia characteristics in finger mark creases: ridge ending, bifurcation, along with short ridge (or dot). Bifurcation is nothing but the point where a solitary ridge segments to two crinkles. Details and trivia remain incredibly imperative to examine fingerprint because no two finger marks are alike.

Two factors come into play for deciding the suitably of finger-print based biometrics than others. These are space complexity and time complexity (Brindha and Vennila, 2011). A bare minimum sized fingerprint stencil has size of bytes more than a few hundred where customary smart-cards possess nonvolatilizable memory of 8K up to 16K approximately. Hence, place intricacy is not a chief crisis for as latter is able to hoard whole finger mark templet. In case of period complication, its processor should be gifted of accomplishing an intact fingerprint toning algorithm instantaneously.

The main aim of this paper is to hide the fingerprint in the person’s photo which can be later printed on his/her Smart card or ID card for identification as well as authentication purpose. In addition, this work mainly focused on the software implementation of hiding the fingerprint in the image for authentication purpose (Yang et al., 2007). To enhance security, prior to embedding, the fingerprint is encrypted (Salem et al., 2011) and then embedded (Amirtharajan and Rayappan, 2012a; b ; c ; d). The fingerprint will be encrypted using an encryption algorithm which will be later discussed. Then the encrypted bits get rooted in covers through Scattered (LSB) embedding. Rather than infixing covert bits in linear fashion as in casual LSB substitution method (Chan and Cheng, 2004), here it is done so in a unique non-linear fashion based on a PRNG which gets generated according to covers’ size.

Cryptography (Schneier, 2007), Steganography (Stefan and Fabin, 2000; Thenmozhi et al., 2012; Zhu et al., 2011; Zhao and Luo, 2012) and Biometrics are blend together to implement a smart Identity Card (ID card) by hiding the fingerprint in the person’s photo which can be later printed on his/her Smart card or ID card for identification as well as authentication purpose. Thus the stored/hidden biometric template (finger print) can be compared with the live template for authentication purpose and the photo is an indication for his/her identity. Steganography (Cheddad et al., 2010), watermarking (Abdulfetah et al., 2010; Zeki et al., 2011) and Cryptography (Salem et al., 2011; Zaidan et al., 2010) play a vital role in providing security (Hmood et al., 2010a, b; Rajagopalan et al., 2012).

Steganography can be broadly classified into spatial (Luo et al., 2011; Mohammad et al., 2011) and frequency domain (Amirtharajan and Rayappan, 2012d; Provos and Honeyman, 2003; Thanikaiselvan et al., 2011a). Another classification is based on the cover object (Bender et al., 1996) like text (Al-Azawi and Fadhil, 2010; Xiang et al., 2011), video (Al-Frajat et al., 2010), audio (Zhu et al., 2011) or in an image (Gutub, 2010; Amirtharajan et al., 2012; Janakiraman et al., 2012a, b; Padmaa et al., 2011; Zaidan et al., 2010; Zanganeh and Ibrahim, 2011; Thanikaiselvan et al., 2011b). In this study, an optimistic effort has been taken to encrypt the finger print through ingenious symmetric key crypto system, later embedded in the photo to build smart ID cards.

Some useful cryptographic system: Since a Simple XOR Cipher with a single key can easily be guessed by the intruders, we have implemented a three key Cryptographic algorithm. The Encryption and Decryption algorithms will be discussed in this section.

Encryption algorithm: Figure 1 shows the encryption algorithm with 3 keys. It consists of 3 different keys say 1, 2 and 3. The first one will be the user defined key i.e., the user has to enter the key he wants to use for encrypting the fingerprint image (INPUT). The other two keys key 2 and key 3 will be the 1-bit left rotated versions of the previous ones. That means key 2 will be the 1-bit left rotated version of key 3 and soon as shown in the above figure.

Fig. 1:	Encryption algorithm with 3 different keys

As per the Figure, the fingerprint image will be first XORed with key 1 and the output of the first XOR Cipher will be again XORed with key 2 and soon. The resulting output will be the encrypted image which was XORed three times using a set of 3 different keys. This is about the encryption algorithm used for encrypting the fingerprint image.

Algorithm: Fingerprint encrypting procedure

Decryption algorithm: The decryption algorithm will be the same as that of the encryption algorithm which we have discussed earlier but in the reverse order as shown in the Fig. 1. The encrypted fingerprint image will be the input to the decryption algorithm and the keys are given in the reverse order.

The same procedure will be repeated for the decryption algorithm but the only change is that here the input to the first XOR cipher will be the encrypted fingerprint image and the key to it will be key 3. The same procedure is repeated as shown in Fig. 1 and the resulting output will be the original fingerprint image.

Algorithm: Fingerprint decrypting procedure

Fig. 2:	Decryption algorithm with 3 different keys

PROPOSED METHOD

LSB is a simplest plus effective technique in data embedding technique. It directly embeds the secret bits in LSB of cover image. Even it introduces distortion when the embedded bits are more than three. This classical paradigm buries the data within cover via., engaging the bit stream of the message to substitute that of the cover’s LSB sequentially. Image steganography intends for preserving numerical traits of congregating image in order to withstand or foil steganalysis. Nevertheless, LSB routines pioneer a little deformation in geometric properties of carrier signal or image to vindicate manoeuvring by means of steganalysis runs. To facilitate this susceptibility, a modus operandi as shown in Fig. 3 which performs scattered LSB embedding in addition to preservation of cover images’ histogram is proposed.

In this technique, the secret is going to be embedded in a non-linear fashion based on Pseudo Random Sequence Generation (PRNG) contrary to the classical LSB steganography, where the confidential data is embedded inside a carrier file in a linear fashion. The pseudo random number is generated depending upon the size of the carrier. This process will be same as that of the classical LSB substitution method but the secret will be embedded in the pixels as decided by the pseudo random number. This pseudo random number must be produced during the extraction process.

As discussed earlier, the main aim of this study is to hide the fingerprint of a person in his/her own photo. For embedding the fingerprint in the cover image, a new Steganographic technique Scattered LSB embedding has been proposed.

Fig. 3:	Block diagram of proposed method

This model suggests a method, where the surreptitious information gets rooted in the non-linear fashion as discussed above. As an improvement, prior to embedding, the bits are encrypted and then embedded.

ENCRYPTION AND DATA HIDING

Algorithm: Embedding process

Algorithm: Extracting/retrieving process

Fig. 4:	Flowchart illustrating the embedding process along with encryption block

Obtained fingerprint output from retrieval process and the original fingerprint can be compared later:

The final result will be the stego image containing the encrypted fingerprint image in the cover image. This stego image can later be printed on ID cards or Smart card for authentication as well as identification purpose. The encrypted fingerprint image is retrieved back using this defined extracting process and decrypted using decryption algorithm. Then the decrypted fingerprint image can be compared with the original fingerprint image.