A Novel Method for Compression And Protection of Biometric Images

Abstract

ABSTRACT newlineIn present scenario biometric systems are widely used in various applications, but the recent newlinebiometric technology faces security issues like stored biometric template attack, where the newlineoriginal data is tampered such that it can t be retrieved and the storage space issues like newlinerequirement of large storage space for biometric data of large population. To avoid such newlinesecurity and compression issues a novel method is proposed. newlineThe proposed method is a combination of multiple novel algorithms like advanced preprocessing newlineand feature extraction algorithms for biometric images, new robbust algorithms newlinefor cryptography and watermarking to reduce security and compression issues. Fingerprint newlineand iris are the biometric characteristics used in this research. The pre-processing method for newlinefingerprint includes Gray Scale Conversion, Gabor Filtering, Thresholding, Contrast Limited newlineAdaptive Histogram Equalization (CLAHE) and Morphological Operation thinning. From newlinethe pre-processed fingerprint image features are extracted using Point Detection Algorithm newline(PDA). The pre-processing methods for iris include Gray Scale Conversion, Gabor Filtering, newlineCanny Edge Detection, Thresholding, Connected Component Labeling using Modified Classical newlineAlgorithm, Pupil Hole Filling and finally Centroid Detection. Features extracted are newlineembedded in watermark image by means of Grayscale Deformation Algorithm (GDA). The newlinegrayscale deformed watermark image along with two pseudorandom sequences generated newlineusing Cryptographically Secure Pseudo-Random Number Generator (CSPRNG) working on newlineCSPRNG-2 Algorithm aided with a cryptographic key is used in the proposed PeppyWavelet newlineWatermarking Algorithm (PWWA). PWWA is based on DiscreteWavelet Transform (DWT) newlineand Discrete Cosine Transform (DCT) watermarking techniques which combines the virtues newlineof CSPRNG-2 and GDA to provide efficient watermarking. The watermark image is recovered newlineby reverse embedding procedure utilizing the proposed Peppy Wavelet Watermark newlineExtraction Algorithm (PWWEA). From that the biometric