ABSTRACT
The main Objective of the project is to generate a biometric based cryptographic key from facial features. Existing asymmetric encryption algorithms require the storage of the secret private key. Stored keys are often protected by poorly selected user passwords that can either be guessed or obtained through brute force attacks. Combining biometrics with cryptography is seen as a possible solution but any biometric cryptosystem must be able to overcome small variations present between different acquisitions of the same biometric in order to produce consistent keys. A new method which uses entropy based feature extraction process coupled with Reed-Solomon error correcting codes that can generate deterministic bit-sequences from the output of an iterative one-way transform. The technique is evaluated using face data and is shown to reliably produce keys of suitable length for 128-bit Advanced Encryption Standard (AES).
Communications advancements in recent decades have led to an increased volume of digital data traveling through publicly shared media. This has led to the rapid development of cryptographic techniques such as AES and public some key architecture. Although keys of sufficient length are strong against both brute force and factorization attacks they still suffer from weaknesses due to insecure key protection by user selected passwords. The limitations of passwords are well documented they are simple and can be easily guessed or obtained using social engineering techniques. They are often written down and stored in an insecure location, can be shared between users, and cannot provide a guarantee of non-repudiation. Furthermore, most people tend to use the same password for a wide range of applications and as a result the compromise of one system leads to the compromise of many others. In recent years researchers have turned towards merging biometrics with cryptography as a means to improve overall security by eliminating the need for key storage using passwords. During the last decade biometrics has become commonly used for identifying individuals. The success of its application in user authentication has indicated that many advantages could be gained by incorporating biometrics with cryptography. A biometric is an inherent physical or behavioural characteristic of an individual such as their voice, face, and fingerprint or keystroke dynamics.
Biometrics, in contrast to passwords, cannot be forgotten, are difficult to copy or forge, impossible to share and offer more security then a common eight character password. The principal drawback of a biometric is that it is not precise like a password or cryptographic key. Limitations of acquisition technology and the inherent changes in the biometric (such as pose and expression for faces) and environmental conditions (such as lighting) lead to variations in each sample of the same biometric. For example, although an iris is considered to be the most accurate of biometrics, there can be up to 30% variation between two different images of the same iris. It is the primary challenges of all biometric cryptosystems to overcome this variation whilst harnessing the advantages of biometrics in order to improve the security of encryption keys. Another challenge stems from the permanence of a biometric. Apart from physical damage, fingerprints or iris remain largely unchanged throughout a person’s life. This is a desired property in most applications of biometrics but in cryptography this is a weakness. Cryptographic keys need to be (and they often are) revoked or changed both proactively as a measure to increase security and reactively as response to key compromise. Most proposed schemes ultimately come down to the protection of an existing cryptographic key with biometric information. While the existing key can be changed the biometric used to secure it cannot and this shortcoming is often neglected.
A cryptographic key generation using face biometric was implemented and verified. A method which uses an entropy based feature extraction process coupled with Reed-Solomon error correcting codes that can generate deterministic bit-sequences was implemented. Combining biometrics with cryptography is a possible solution but any biometric cryptosystem must be able to overcome small variations present between different acquisitions of the same biometric in order to produce consistent keys. Hence the biometric key obtained from the face as been combined with the cryptographic key and the look-up table was generated in order to obtain the cryptographic key while decoding. The method is also flexible, the bio-keys used to protect the cryptographic key can be changed and revoked and is a significant feature not possessed by other methods. The method can be modified to protect keys of increasing length by either increasing the size of bio keys through performing more rotations/iterations of the bi spectral transform or by changing the RS encoding scheme used. A standard 128bit AES algorithm is used here and the obtained cryptographic key is verified using that 128bit AES algorithm
authors: PAVANI (Btech RVR&JC COLLEGE OF ENGG.) S.TILAK (Btech RVR&JC COLLEGE OF ENGG.) VINOD KUMAR (Btech RVR&JC COLLEGE OF ENGG.) KRANTHI KUMAR (Btech RVR&JC COLLEGE OF ENGG.) I sincerely thank authors,the students of "RVR&JC COLLEGE OF ENGG" for their great support. my special thanks to v.v.m.m.rao
No comments:
Post a Comment