Image Encryption of Biometric Characteristics Using Chaotic System

Nowadays ,due to the rapid development of multimedia and digital technology the transmission over public channel become more and more. Also the issues with the confidentiality and integrity of data increases in an alarming rate. So that the personal identification information or biometric signatures requires a security technique to protect from the any of the unauthorized access. Biometric signatures such as fingerprint , iris , palm print, gestures, handwritten signature and hand geometry have to be considered as an efficient tool for establishing the identity of an individual. Because they contain the unique characteristics of a person. So we go for the proposed biometric information security system using a three fold encryption technique. Which utilizes an orthogonal coding and multiplexing technique, direct LSB replacement steganography technique and finally a chaotic encryption method. Multiple biometric signatures are encoded and then multiplexed together in the form of a single image using orthogonal encoding and multiplexing.The encoded image is embedded in to the colour cover image ,which is decomposed in to three plane namely red ,green ,blue respectively. In steganography technique, the multiplexed image pixel is used to replace the LSB of corresponding colour cover image pixel for information hiding purpose. Finally, the stego image is encrypted using a new one dimensional chaotic system based MRJTC encryption technique.The objective of this paper is to Protect the Biometric Information from any unauthorized access and also reduces the vulnerability of an intruder to retrieving any information through any steganalysis attack .In addition to this technique can accommodate a number of different biometric information in the same cover image while maintaining the negligible amount of distortion when compared with original colour cover image. The encrypted images are random, non repeated and unpredictable. Chaotic system based encryption have excellent diffusion and confusion properties and can resist the any plaintext attack. Performance of the proposed technique is investigated through matlab simulation using various biometric signatures and colour cover images.